Professional Documents
Culture Documents
On Measuring Student-Teacher Relationships - Sorting Out Predictor
On Measuring Student-Teacher Relationships - Sorting Out Predictor
Fall 2015
Recommended Citation
Barch, Jon Craig. "On Measuring Student-Teacher Relationships: Sorting Out Predictors, Outcomes, And Schematic Structure Of
Students’ Internal Relationship Representations." PhD (Doctor of Philosophy) thesis, University of Iowa, 2015.
https://doi.org/10.17077/etd.y852ur5n
by
December 2015
2015
CERTIFICATE OF APPROVAL
____________________________
PH.D. THESIS
_________________
____________________________________________
David Lohman
____________________________________________
Joyce Moore
____________________________________________
Walter Vispoel
____________________________________________
Paul Windschitl
To my mother, Joan Barch and my father, James Barch for providing me with the
autonomy I required as a child, the structure that guided my strivings, and the love
that provided me confidence for a lifetime. To Edward Neimi for allowing me to
remain a student at Northern Michigan University when many would have sent me
home. To David Bonsall and Rachel Harris for seeing the potential for socially
responsible leadership development in me, despite my conduct record. To Dr. Alan
Beauchamp for introducing psychology to me through an Advanced Personality
Theory course that I had no prerequisite knowledge for but enjoyed tremendously.
To Dr. Cynthia Prosen for showing me how interested and excited one could be
when teaching psychology to others, an experience which inspired me toward a
career in academia. To Dr. Shelia Burns who built in me a strong foundation for
understanding statistics; carpe datum. To Dr. Bradley Olson who introduced me to
Social Psychology, life as a psychological researcher, and Self-Determination
Theory. To Dr. Edward Deci and Dr. Richard Ryan, the architects of Self-
Determination Theory, as they have provided me boundless inspiration through
their prolific writings. To Dr. Johnmarshall Reeve who agreed to bring me to the
University of Iowa as one of his graduate students. To Dr. David Lohman who was
truly the academic mentor and friend that I needed at the University of Iowa to
retain hope in completing this project. To Dr. Kathy Schuh who graciously agreed
to support me across the finish line. And with more gratitude, emphasis, and import
than all previous remarks, to my wife Maggie, daughters Lenora and Fiona, and son
James for encouraging, supporting, tolerating, and providing me with a sense of
purpose over the last ten years that I have been working on this project.
ii
“It is, in fact, nothing short of a miracle that the modern methods of instruction have
not yet entirely strangled the holy curiosity of inquiry; for this delicate little plant,
aside from stimulation, stands mainly in need of freedom. Without this it goes to
wrack and ruin without fail. It is a very grave mistake to think that the enjoyment of
seeing and searching can be promoted by means of coercion and a sense of duty.”
Albert Einstein
As quoted in Einstein and the Poet
by William Hermanns in 1983
iii
ACKNOWLEDGEMENTS
Let’s be real, I would have never completed this dissertation if Dr. Christine
Michigan University, would have not insisted that I finish during one of her
dissertation without the friendly, but frequent and persistent requests for progress
provided a great deal of guidance and editorial assistance at the dissertation proposal
iv
ABSTRACT
theoretical conceptualizations of the relationships and how they measure them are
often quite different. This study provides empirical insights for both measurement
representations.
examined as well.
The hypothesized model was not supported by the data. The study
acceptable fit. The study provided evidence as to which of the 170 items from the
quality. The study exemplified the method effect dangers of negative item wording.
v
Finally, the study provided strong evidence for conceptualizing student-teacher
vi
PUBLIC ABSTRACT
This is because there are several different theories that researchers reference when
sometimes they differ drastically in their content. This study reviews 14 of the
questionnaire items into one survey, and gathers data from 628 students. The
and further our understanding of how students’ think about their relations with
teachers.
A new model was proposed, but the data did not support it. The study did
allow comparison about how useful each of the 14 questionnaires is. The study
reminds us that multiple models can fit the same data. The study examined which
items from each of the questionnaires most closely measure the core of student-
teacher relationship quality. The study provides an example of how switching the
results. Finally the study provided strong evidence for thinking about student-
a multi-dimensional construct.
vii
TABLE OF CONTENTS
Student-Teacher Relationships........................................................................................ 1
Purpose .......................................................................................................................... 50
Participants .................................................................................................................... 50
Instrumentation.............................................................................................................. 52
viii
Procedure ....................................................................................................................... 68
Hypotheses .................................................................................................................... 70
Achievement.................................................................................................................. 96
FACTORS........................................................................................................................178
ix
APPENDIX D. TEXT FOR INVITATION PARTICIPATE EMAIL ............................188
ASSIGNMENT ................................................................................................................192
DELETION ......................................................................................................................200
DELETION ......................................................................................................................204
REFERENCES ................................................................................................................206
x
LIST OF TABLES
outcomes ............................................................................................................................10
Table 3. CFA model fit statistics for the fourteen original scales .................................... 80
Table 5. SMC estimates for Student Classroom Environment Measure items ................ 82
items .................................................................................................................................. 83
Table 7. SMC estimates for Inventory of Parent and Peer Attachments items................. 85
Table 8. SMC estimates for School Utilization & Emulation items ................................. 85
Table 9. SMC estimates for Research Assessment Package for Schools items................ 87
Table 10. SMC estimates for Teacher as Social Context items ........................................ 87
Table 12. SMC estimates for Emotional Reliance Questionnaire items ........................... 90
Table 13. SMC estimates for Basic Need Satisfaction in relationships items .................. 90
Table 14. SMC estimates for Need for Relatedness Scale items ...................................... 91
Table 15. SMC estimates for Student-Teacher Relationship Scale items ........................ 92
Table 16. SMC estimates for Young Children's Appraisals of Teacher Support items .... 93
Table 17. SMC estimates for Psychological Sense of School Membership items ........... 96
Table 18. Regression results for each original scale with internalization as the
xi
Table 19. Regression results for each original scale with engagement as the
Table 20. Regression results for each original scale with academic achievement as
Table 21. Items deleted from predictor factors in hypothesized model due to low
Table 22. Items deleted from predictor factors in hypothesized model due to high
Table 23. Five highest error covariance modification indices for items in
Table 24. Items deleted from perceptions of teacher measurement model due to low
Table 25. Five highest error covariance modification indices for items in
Table 26. Four highest error covariance modification indices for items in
Table 27. Seven remaining allowable error covariance modification indices greater
Table 28. Items deleted from the outcome variables’ measurement model due to low
Table 29. Items deleted from outcome factors in hypothesized model due to high
xii
Table 30. Regression weight modification indices for treating academic
Table 31. Four highest error covariance modification indices for the Full model as
Table 32. Parameter estimates for factor loadings and path coefficients in the final
Table 33. Bivariate correlations among subscale scores for predictor factors,
relational schema factors, and outcome factors in modified full structural model ......... 134
Table 34. Variance inflation factor statistics for all predictor factors, relational
schema factors, and outcome factors in modified full structural model when
Table 35. Factor loadings of items retained for the three factor solution after items
Table 36. Factor loadings of all items for the single-factor solution .............................. 139
Table 37. Comparison of internal reliability from original publication and this study
xiii
LIST OF FIGURES
items ......................................................................................................................... 81
items ......................................................................................................................... 88
items ......................................................................................................................... 89
xiv
Figure 12. Standardized CFA parameter estimates of Basic Need Satisfaction
added to the C-STRI for examining predictive validity of original scales .............. 98
original scales........................................................................................................... 99
Figure 21. Latent variable covariance and factor loading parameter estimates
xv
Figure 22. Standardized parameter estimates for 3-factor model of predictors
order factor model of perceptions of teacher with all items included.................... 113
of teacher factor model with items removed and error covariance added ............. 116
factor model of Perceptions of Self with all items included .................................. 117
of Self factor model with items removed and error covariance added .................. 120
factor model of Interpersonal Scripts with all items included ............................... 121
Figure 31. Full structural equation model with the hypothesized latent variable
xvi
Figure 32. Full structural equation model with unknown parameters reduced by
calculating subscale scores for several factors from the modified measurement
Figure 33. Full model using subscale scores and treating academic achievement
Figure 34. Full model using subscale scores and treating academic achievement
Figure 35. Forward portion of the principal components analysis scree plot ........ 136
xvii
PREFACE
experience of many youth these days, particularly those in secondary schools. Upon
beyond for that matter, I realize that the interest, enjoyment, and effort I personally
put into learning was consistently tied to the perceptions I had about the quality of
relationships with enough descriptive detail that I might be able to assist future
I also feel compelled to state up-front that this thesis is in no way intended to
refute the validity, reliability, or general usefulness of any of the existing instruments
reviewed throughout the project. Rather, the motivation behind this project is to
build on the great wealth of knowledge and tools that exist in the literature about
xviii
CHAPTER I. INTRODUCTION
they have long been at the core of a broad spectrum of psychological research. It has
even been suggested that perceived relatedness to others is a psychological need and in
order to achieve optimal physical and psychological functioning human beings must
perceive themselves as having high quality relationships within a preferred social group
(Baumeister & Leary, 1995; House, Landis, & Umberson, 1988). Even people who
report not needing close relationships with others show large increases in positive affect
when they learn that others accept them or that they will have interpersonal success in the
future (Carvallo & Gabriel, 2006). The idea that humans require strong emotional bonds
with other humans to be healthy is not new; rather, it has been a major component of
several classic personality theories (e.g., Fairbairn, 1952; Freud, 1930; Fromm, 1956;
Horney, 1945; Maslow, 1943; Rogers, 1951). The prominence of these as well as more
contemporary theories (e.g., Bowlby, 1973; Deci & Ryan, 1985) has fueled volumes of
Student-Teacher Relationships
One set of interpersonal relationships that most children experience are with the
teachers they have throughout their lives. The quality of these student-teacher
relationships can have a profound impact on a child, building a cognitive foundation from
concluded such relationships have significant importance for many social and cognitive
developmental outcomes (e.g., Brophy, 1998; Davis, 2001; Goodenow, 1992; McCallum
1
& Bracken, 1993; Oldfather & Dahl, 1994; Pianta, 1999; Ryan, Connell, & Deci, 1985;
Wentzel, 1999). For example, Davis, (2001) reviewed evidence that student-teacher
the classroom and social competence. Goodenow (1992) discussed evidence that
achievement. Again, this research and more shows us how developmentally important
these relationships can be. As with research in any field, the quality of empirical research
ultimately improve the lives of children who experience them. Accordingly, this
Statement of Problem
diverse. This is likely due to the researchers coming from a variety of training
overlap does exist among them. However, the equally extant inconsistencies leave many
questions about what a high quality student-teacher relationship actually is, how many
distinct factors are appropriate to consider when describing the relationship, what
psychological mechanisms are responsible for its association with positive social and
2
cognitive outcomes, and, perhaps most importantly for good science, how the construct
measures include items that seem to be inputs to the relationship, others include items
that could arguably be considered outcomes of the relationship, and others measure
The diversity of these tools creates confusion about which scores should be considered
predictor or outcome variables of the relationship and which scores should be included in
the measurement model of a latent variable representing aspects of the relationship itself.
and can leave readers questioning what aspects of the relationship are important and why.
cognitive psychology paradigm guiding relationship perception research and theory may
experience, would be a great place to look. Baldwin (1992) expertly synthesized a broad
3
schema theory as a guide. Likewise, a process of blending and evaluating the many
Educational psychologists have not yet attempted such cross perspective theory
literature often pull findings from studies across perspectives, apparently making an
assumption that various researchers are all studying the same concept and measuring the
Purpose of Study
The purpose of this study was to refine the conceptualization and measurement of
teachers' relationships with their students. The many ways educational researchers define
and measure student-teacher relationships were examined to compare and contrast their
teacher relationship perceptions was gathered using instruments from all the perspectives
insights for synthesizing the existing perspectives into a single theoretical framework.
relationships were collected. The process of gathering these instruments is laid out in the
literature review chapter. The items from these instruments were combined into a single
survey, which was administered to a large sample of college student respondents. When
4
experienced with a particular high school teacher, in the past tense. Instrument items
were only adjusted as required to grammatically correspond with this request. Item
adjustments are explained further in the methods chapter. The internal consistency of
items was examined as grouped by the original scale they were taken from. Baldwin’s
respondents were randomly assigned to reflect on experiences with their most recent high
school science, math, or English teacher. Exploratory factor analysis was also used to
hypotheses were tested: 1) when examined separately from other instruments’ items, the
items from each of the original instruments will demonstrate acceptable internal
consistency and fit to their instrument’s expected structural model; 2) calculated scale
scores from each of the original measurement tools will demonstrate acceptable
predictive validity in the expected direction as related to each of the student outcomes
measurement tools taken together will fit well into a new 13-factor structure; and 4) the
new factors will fit well into a relational schema theory driven structural model.
5
Importance of Study
providing students with knowledge, skills, and attitudes deemed important for successful
quality influences student motivation and engagement in the learning process, which
internalization of desired attitudes (Brophy, 1998; Davis, 2001; Pianta, 1999; Ryan et al.,
of relationship quality. This affords educators the ability to facilitate the development of
relationships with students in desired ways and ultimately improve student learning and
development. This study is important because it provides empirical evidence for both
relationship representations.
The past has demonstrated how this kind of cross paradigm integration can be
fruitful. It is usually the case that each paradigm has bits of truth to offer, yet too often
researchers work within one paradigm and fail to explore any synthesis of alternative
interpersonal relationship theorists, for nearly a century, theorists in the study of human
6
intelligence proposed and studied a multitude of models representing intelligence.
Providing a basis for the method of measurement integration utilized by this study of
on several intelligence tests simultaneously provided empirical support for the integrated
theories by Horn and Cattell (1966) as did Carroll’s (1993) extensive factor analytic study
of multiple models and Horn’s (1994) integrative factor analytic study. Similarly, in
personality research, the work of McCrae and Costa (1987) using multiple measures of
personality taxonomies and factor analyses paved the way for eventual widespread
acceptance of the Big 5 theory (Costa & McCrae, 1992). Unfortunately, contemporary
divided to a great extent. These examples of critical progress in the fields of intelligence
and personality research alike demonstrate how consideration of concepts and ideas from
a broad range of research perspectives can be helpful and the integrative process fruitful.
Taking a page from this history, the present study was carried out to explore a newly
7
CHAPTER II. LITERATURE REVIEW
Five prominent perspectives are reviewed in the subsequent sections of this chapter;
the process used to identify them and the parameters of inclusion in this review. To
begin, because this project is focused on measurement integration, broad searches using
PsycINFO and ERIC were conducted to identify the tools used in quantitative research on
“questionnaire,” “instrument,” or “scale” along with all three terms “student,” “teacher,”
and “relationship.” After finding three other terms, “relatedness,” “involvement,” and
further searches were added substituting each of these terms for “relationship,” one at a
time. All searches were conducted with restrictions to include only journal articles or
book chapters and to exclude anything prior to 1980. The resulting studies were
manually sorted into two categories, 1) review articles and 2) original research that
included a quantitative measure of the student-teacher relationship; all other articles were
discarded. The review articles were used to further locate additional, original-research
sections of the original-research articles were also used to reverse mine citations of
8
quantitative research on student-teacher relationships not yet identified as well. Next, the
mass of original-research articles was sorted by the instrument used to measure the
student-teacher relationship. For example, all studies that measured the relationship with
the Student-Teacher Relationship Scale (Pianta & Nimetz, 1991) were placed into a
group. The next step was to search PsycINFO and ERIC several more times using the
names/titles of the instruments identified in the sorting process to identify any studies that
may have used those measures but had not yet been identified. After this, studies were
discarded if they used a measure that was not also used in any other studies or if the
measure described was not based on any identified psychological theory on human
relationships. What remained were fourteen instrument groupings. Because this study is
intended to afford both measurement integration and theory integration, the instrument
groupings were organized into five higher-order groupings according to the foundational
relationship theory referenced by the researchers who first developed the instrument. In
relatedness need support (Deci & Ryan, 1985) when developing their instrument, were
and dependency, and 5) sense of belonging. In the following five sections of the paper,
these relationship perspectives are briefly reviewed, the relationship measurement tools
associated with them are examined, details from a few major studies are shared, and a
listing of outcomes found to be significantly related to each measure are laid out. Table 1
9
Table 1. Summary of relationship measurement scales, associated constructs, and outcomes
Reference Relationship Scale Title Relevant # of Coefficient Sample Item(s) Associated Outcomes
Construct Subscales Items Alpha
Johnson, Pedagogical Classroom Teacher social 4 .80 My teacher cares about and Academic interest, social
Johnson, caring Life Measure support likes me as a person. responsibility, academic pro-social
Buckman, & behavior, mastery orientation toward
Richards, 1985 Teacher 4 .78 My teacher cares about how learning, academic effort, engagement,
academic much I learn. self-regulation, & academic
support achievement
Feldlaufer, Pedagogical Student Teacher fairness 7 .70 - .75 The teacher cares how we Academic motivation, task valuing, &
Midgley, & caring Classroom & friendliness feel. task interest/enjoyment
Eccles, 1988 Environment
Measure
Developmental Caring Developmental Classroom 10 .82 My class is like a family. academic motivation, intrinsic
Studies Center, school Studies Center supportiveness motivation, task orientation, school
2002 community Student In my class the teacher and enjoyment, prosocial attitudes, and
Questionnaire Meaningful 10 .80 students decide together educational attainment expectations,
participation what the rules will be. less ego orientation, & work avoidance
Armsden & Relatedness Inventory of Trust 10 .91 My __ understands me. Positive coping at school, self-
Greenberg, Parent & Peer regulated learning, perceived control,
1987 Attachments Communication 9 .91 I tell my __ about my school engagement
problems and troubles.
Ryan, Stiller, & Relatedness School NA 3 .66 I can usually rely on my Positive coping at school, self-
Lynch, 1994 Utilization teacher when I have regulated learning, perceived control,
problems at school. school engagement
10
Table 1. Continued
Institute for Relatedness Research Teacher 4 .71 When I am with my teacher School emotional and behavioral
Research and Assessment Emotional I feel happy. engagement
Reform in Package for Security
Education, 1998 Schools
Belmont, Relatedness Teacher as Affection 3 .71 My teacher likes me. School engagement, internalization of
Skinner, Social Context task importance, autonomous
Wellborn, Attunement 3 .54 My teacher knows me well. motivation
Connell, 1988
Dedication of 2 My teacher spends time
Resources with me.
Furrer & Relatedness Sense of NA 4 .79 When I am with my teacher Emotional engagement, behavioral
Skinner, 2003 Relatedness I feel accepted. engagement, & help-seeking behavior
Butzel & Ryan, Relatedness Emotional NA 7 .91 - .97 If I were feeling alone or Well-being, mental health, & perceived
1997 Reliance depressed, I would be self-determination
Questionnaire willing to turn to my _____.
LaGuardia, Relatedness Basic Need Relatedness 3 .90 When I am with ____ I feel Fully mediates relationship between
Ryan, Satisfaction in loved and cared about. attachment security and well-being
Couchman, & Relationships
Deci, 2000
Richer & Relatedness Need for Acceptance 5 .89 In my relationship with my Increased vitality, increased
Vallerand, 1998 Relatedness _______, I feel supported. performance
Scale
Intimacy 5 .91 In my relationship with my
_______, I feel as a friend.
11
Table 1. Continued
Pianta & Closeness, Student- NA 16 .85 This student trusts me. Social competence, frustration
Nimetz, 1991 conflict, Teacher tolerance, work habits, self-discipline,
dependency Relationship classroom participation, cooperation,
Scale (STRS) academic competence, school
achievement, less behavior, conduct, &
attention problems, peer aggression,
internalizing, & school avoidance
Mantzicopoulos Closeness, Young Warmth 11 .75 My teacher is my friend. Reading ability, self-control, less
& Neuharth- conflict, Children’s emotionality, school achievement,
Pritchett, 2003 dependency Appraisals of Autonomy 6 .67 My teacher lets me do cooperation, self-control, & less
Teacher activities I want to do. behavior problems
Support
(Y-CATS) Conflict 10 .75 My teacher gets angry with
me.
Goodenow, Sense of Psychological NA 18 .87 Most teachers at (name of task valuing, success expectancies,
1993 belonging Sense of school) are interested in me. academic motivation, positive affect,
School optimism, academic efficacy, intrinsic
Membership goal orientation, school achievement,
less risky behaviors, depression, social
rejection, & behavioral problems
12
Pedagogical caring. Research on student-teacher relationships from this point of
view places the focus on the teacher’s relational style. High quality relationships are
aimed at mutual understanding. They know students’ ability level, have appropriately
high expectations of their students, and they model compassionate behavior toward their
students (Noddings, 1992). Put differently, students seem to thrive in a respectful, home-
like environment (Matzye, 1995); whereas, the lack of caring teachers has been
of Nodding’s (1992) view of the caring teacher, Wentzel (1997) analyzed descriptions of
teachers’ caring and non-caring behaviors provided by 375 eighth-grade students. Caring
teachers were generally described as making an effort to capture student interest, they
Wentzel (1994; 1997; 1998) using the Teacher Social Support and Teacher Academic
Support subscales of the Classroom Life Measure (Johnson & Johnson, 1983; Johnson,
Johnson, Buckman, & Richards, 1985). Johnson and Johnson (1983) report that the two
subscales formed separate factors using a varimax rotation factor analyses. Teacher
Academic Support was assessed with 4 items measuring a student’s belief that the teacher
1) cares about how much he/she learns, 2) likes to see his/her work, 3) likes to help
him/her learn, and 4) wants him/her to do his/her best in schoolwork and showed strong
13
internal consistency (α= .85). Teacher Social Support was assessed with 4 items
measuring a student’s belief that the teacher 1) really cares about him/her, 2) thinks it is
important to be his/her friend, 3) likes him/her as much as other students, and 4) cares
about his/her feelings and showed reasonable internal consistency (α= .68). As such,
many others have used the subscales separately (e.g., Patrick, Kaplan, & Ryan, 2011;
Patrick, Ryan, & Kaplan, 2007; Ryzin, Gravely, & Roseth, 2009). However, it is unclear
how distinct these subscales actually are. Wentzel (1997) found the two sub-scales to be
highly correlated in both sixth (r = .67) and eighth (r = .71) grade students so the
subscales’ items were combined to form a single composite score, which showed strong
internal consistency in both sixth (α= .89) and eighth grade (α= .91). Wentzel and her
colleagues now regularly combine items from both subscales into one construct depicting
how much the student perceives that the teacher cares. Not surprisingly, their data
similarly support the assertion that students with caring or nurturing teachers have
increased academic interest (Wentzel, 1997; 1998), positive social and academic goal
striving (Wentzel, 1994; 1997; Wentzel, Baker, & Russell, 2012), mastery orientation
toward learning (Wentzel, 1997), and greater adherence to classroom norms and rules
support together influence students’ general perception that the teacher cares about them
(Wentzel 1994, 1998; Wentzel, Baker, & Russell, 2012). Support for this single construct
approach is also found in a recent repeated measures study using these same items as part
of a survey administered to 283 secondary school students by Van Ryzin, Gravely, and
Roseth (2009). These researchers modeled how academic autonomy and support in
school can predict engagement in learning, which in turn predicts children’s positive
14
psychological adjustment. As expected, student perceptions of teacher support was an
Gravely, & Roseth, 2009). Of interest to the measurement issue, though, is that these
researchers began by assuming the subscales of teacher academic and social support were
separate constructs, but found them to be one factor after doing principal axis factor
analysis using Promax rotation. When treated as a single construct, Van Ryzin, Gravely,
and Roseth’s (2009) data produced strong internal reliability figures at both measurement
However, in another recent study, Patrick, Ryan, and Kaplan, (2007) examined
predicted classroom engagement and if that relationship was mediated by the students’
motivational beliefs. As with nearly all the studies on this topic, they administered a
battery of surveys to gather the data from the children. Teacher academic and
the researchers tested their proposed model using them as separate constructs as
compared to a model in which the items for both subscales loaded onto a single construct
for teacher support. They found that significant information was gained by keeping the
subscales separate. More importantly, their data provide more evidence that when a
teacher is perceived as more supportive, the student is more likely to use self-regulatory
strategies and engage in more on-task behaviors, which result in greater academic
achievement. Further, their results suggest that this relationship is mediated by the
effects that the teacher supportiveness has on students’ use of mastery goals and
15
perception of academic self-efficacy (Patrick, Ryan, & Kaplan, 2007). Patrick, Kaplan,
and Ryan (2011) used multidimensional scaling with three samples of adolescents (Ns =
537, 537, and 736) as well as structural modeling with a fourth sample (N = 789) to
examine the dimensionality of goal structures and classroom climate including teacher
academic support and teacher social/emotional support. Although, the data suggested
strong correlation between teacher academic support ratings and teacher social/emotional
support ratings (r = .67-.77), the spatially plotted data supported the multidimensional
this section on pedagogical caring, Feldlaufer, Midgley, and Eccles (1988) developed
their own 6-item scale to measure students’ perceptions of teachers as supportive, caring,
friendly, and fair in order to examine the differences in students’ perceptions of teachers
before (α= .70) and after (α= .75) the transition from elementary school to middle school.
This subscale, titled Techer—Unfair/Unfriendly, was part of a larger measure called the
Student Classroom Environment Measure and its six items asked students if teachers
cared about how they feel, were friendly to them, treated boys and girls differently,
graded work fairly, treated some kids better than others, and criticized poor work
(Feldlaufer et al., 1988). Midgley, Feldlaufer, and Eccles (1989) used this measure again
changes from elementary school to middle school in relation to the change in perceptions
of their teachers. These data suggested main effects of perceived teacher support on math
16
F(3,1300) = 35.59, p < .0001, as well as, the two-way interaction between change in
F(3,1300) = 21.80, p < .0001 and math usefulness/importance, F(3,1300) = 16.41, p <
.0001. Before moving on, it is important to note the similarity between this
discussed. Not surprisingly, many of the positive outcomes found by researchers using
(Battistich, Solomon, Kim, Watson, & Schaps, 1995; Solomon, Battistich, Kim, &
both definition and measurement than school belongingness. Further, although both
concepts are relevant to this project, and thus both are included in this review, the two
concepts come from different research traditions. Therefore, they have been intentionally
community theorize that students who experience a sense of school community “will feel
strongly attached to the community and that this attachment will lead them to feel
personally committed to the values and goals the community promotes, particularly if
those values and goals are clear and jointly held by the community members” (Solomon,
Battistich, Watson, Schaps, & Lewis, 2000, p.5). Though its measurement has undergone
some revision, it still consists of two main elements: 1) students’ perceptions of the
17
school as supportive both emotionally and cognitively and 2) students’ beliefs that they
have an influential role in decision making. Originally, it consisted of 24 items split into
two subscales accordingly (Solomon et al., 1997). Most recently these subscales, titled
Classroom Supportiveness and Meaningful Participation each have 10 items and exist
within a larger instrument called the Student Questionnaire, which is authored and
(CDP) (Solomon et al., 2000; Solomon, Watson, Delucchi, Schaps, & Battistich, 1988).
committed to the norms and values emphasized by the school (Battistich et al., 1995;
associated with academic motivation (Solomon et al., 1996). Sense of community has
also been positively associated with intrinsic motivation and task orientation, but
negatively associated with ego orientation and work avoidance (Battistich et al., 1995).
also higher for students perceiving school as a community, an effect which is especially
strong for impoverished students (Battistich, Solomon, Watson, & Schaps, 1997).
Additionally, sense of community has been positively associated with students’ district
achievement test scores; however, this relationship diminished when students’ poverty
18
When discussing the importance of students’ experience of a strong sense of
community in school, this cohort of researchers often cite the satisfaction of the three
basic psychological needs for autonomy, competence, and relatedness that are explicated
by self-determination theory (Deci & Ryan, 1985). While they claim to be extending the
satisfying the three basic needs in one’s community (Solomon et al., 1997), it is more of
engagement in the social world (Deci & Ryan, 1985; Ryan & Deci, 2000).
As such, the most recent reports from the Child Development Project (Battistich
et al., 2000; Solomon et al., 2000), which measure students’ perceptions of school
nice link to the next higher-order classification, which covers self-determination theory
autonomy and relatedness need supports along with their perceptions of supportive
psychological needs critical for optimal human functioning and development (Deci &
Ryan, 1985). Grounded in attachment theory (Bowlby, 1973) and object relations theory
19
(Behrends & Blatt, 1985; Fairbairn, 1952; Winnicott, 1965), the need for relatedness is
schemas derived from, and used on-the-fly for socializing interactions with others (Ryan,
Avery, & Grolnick, 1985). These schemas should be measurable through accessing an
1985; Crittenden, 1990; Lynch & Cicchetti, 1991), which contains both knowledge and
empirical work as compared to research on the needs for perceived competence and
autonomy (e.g., Deci & Ryan, 1991; Ryan & Deci, 2000; Skinner, Furrer, Marchand, &
lacks definitional clarity and measurement consistency, which can be inferred from the
variety of ways students’ perceptions of relatedness support from their teachers have been
measured.
For example, using the self-determination theory framework, Ryan, Stiller, and
Lynch (1994) measured 606 middle school students’ relationships with teachers, parents,
and peers and among other findings they determined that students’ relationships with
teachers are strong predictors of their academic motivation. In this study, student-teacher
20
coefficient alphas of .55, .80, .66, and .84 respectively (Ryan et al., 1994). The first two
(IAA) (Greenberg, 1982 as cited in Ryan et al., 1994). This inventory has since been
revised. It now has 25 items split into three subscales trust, communication, and
alienation and is called the Inventory of Parent and Peer Attachment (Armsden &
Greenberg, 1987; Pace, Martini, & Zavattini, 2011). Ryan et al. (1994) substituted the
word “teacher” for “mother/father” in the parent items to make it an inventory of teacher
attachment. For their other relationship representation variables, they measured school
utilization with 3 items focused on whether or not the student used the teacher when
encountering academic problems, and they used 5 items that concerned student’s desire to
be liked and be seen as similar to their teacher to measure emulation (Ryan et al., 1994).
theory conceptually, Lynch and Cichetti (1992; 1997) used a 17-item relatedness
questionnaire adapted from Wellborn and Connell’s 1987 unpublished Manual for the
Rochester Assessment Package for Schools (as cited in Lynch & Cichetti, 1992). Lynch
and Cichetti (1992) used the questionnaire to measure differences in seven to thirteen
relationships with teachers and parents determining that teachers can serve as an
attachment relationships with. Lynch and Cichetti (1997) used this instrument again to
assess 1,226 second through eighth grade students’ relationship patterns with their
such as the finding that relationship quality with teachers and parents declines during
21
these years while relationships with peers improves. This questionnaire has two
emotional quality items use a 4-point scale to assess children’s emotional experience;
specifically, the items assess the degree students’ feel relaxed, ignored, happy, mad,
bored, important, unhappy, scared, safe, and sad when they are with the teacher. The
seven psychological proximity seeking items use a 4-point scale to assess student’s
desiring more attention from, to spend more time with, to be better known by, to have
feelings perceived by, to be closer to, to enjoy more time with, and to be better able to
talk about things with their teacher. The optimal relationship is defined by high
emotional quality scores (high positive emotion) and low psychological proximity
seeking scores (satisfied with existing degrees of closeness). In the latest version of this
instrument, put out in 1998 by the Institute for Research and Reform in Education, the
emotional quality subscale has been reduced to 4 items and is now called teacher
emotional security; whereas, the psychological proximity seeking subscale was dropped
due to low reliabilities and lack of consistent associations with indices of student
performance and adjustment (IRRE, 1998). The reported reliability sample consisted of
2429 diverse middle school students with a teacher emotional security subscale
Using yet another measure, but again advocating the self-determination theory
perspective, Skinner and Belmont (1993) operationalized support for the relatedness need
children that validated their model suggesting need support influences student
engagement, which in turn influences student adjustment and achievement. Further, this
22
study provided strong evidence for the reciprocal effects of student engagement back on
teachers’ providing of need support in the classroom (Skinner & Belmont, 1993).
Skinner and Belmont (1993) measured teacher involvement using a shortened 8-item
version (alpha = .79) of their Teacher as Social Context (TASC) scale (Belmont, Skinner,
Wellborn, & Connell, 1988) with 2 items tapping teacher’s affection (liking and
items on dedication of resources (aid, time, energy), and 2 items regarding dependability
(available in case of need). The TASC also has a longer, 14-item version (alpha = .83)
with 3, 3, 2, and 6 items for each subscale respectively (Belmont et al., 1988). Bao and
Lam (2008) used the TASC to examine the relationships between perceptions of
high, which provides support for the contention that relatedness facilitates internalization
and allows students to feel autonomous when doing teacher-determined tasks (Bao &
Lam, 2008). A more recent study used the TASC as a basis for the development of a
and 12 Indonesian classrooms (Maulana, Opdenakker, Stroet, & Bosker, 2013). This
study revealed that during the first year of secondary education teacher involvement
decreased over the year and that students accordingly reported less autonomous academic
Meanwhile, Furrer and Skinner (2003), Marchand and Skinner (2007), as well as
23
relatedness to their teacher using a different four-item scale: 1) When I’m with my
teacher I feel accepted, 2) like someone special, 3) ignored, and 4) unimportant with good
internal consistency (alphas = .79-.93). Furrer and Skinner (2003) found this measure of
relatedness to predict changes in 641 third to sixth grade children’s engagement across a
school year over and above the effects of perceived control. Marchand and Skinner
autonomy supportiveness, in their study of 765 third to sixth grade children. Skinner,
year and found scores on this measure of relatedness during the fall semester to be an
Butzel and Ryan (1997) explain the measurement of another aspect of relatedness
experiences with others, both positive and negative, calling their measure the Emotional
Reliance Questionnaire (ERQ). Specifically, the ERQ asks if the child would be willing
to turn to a specific adult (concerning this project, their teacher) when 1) depressed, 2)
consistency scores were not reported for the scale; however, scores on this measure were
associated with less depression and greater well-being (Butzel and Ryan, 1997).
24
Formulating yet another measure of relatedness need satisfaction, founding self-
determination theorists Richard Ryan and Edward Deci collaborated with colleagues to
develop and use a Basic Need Satisfaction in Relationships scale (LaGuardia, Ryan,
Couchman, & Deci, 2000), which has 3 relatedness subscale items that when applied to
student-teacher relationships share the stem “when I am with my teacher” and finish with
1) I feel loved and cared about, 2) I often feel a lot of distance in our relationship, and 3) I
feel a lot of closeness and intimacy. Although they did not examine relationships
between students and teachers, it was found that basic need satisfaction measured in this
way (alpha = .92) fully mediated the relationship between attachment security (with
mother, father, romantic partner, roommate, and other significant adult) and well-being
Vallerand created another relatedness assessment (Richer & Vallerand, 1998). This scale
consists of 10 items split into the two, 5-item subscales titled acceptance and intimacy.
This scale can be used for any relationship by substituting the blank with coach, mother,
supervisor, teacher, etc. in the item stem that reads “In my relationship with my ______, I
feel…” The 5 items in the acceptance scale (alpha = .89) read 1) supported, 2)
understood, 3) listened to, 4) valued, and 5) safe. The 5 items in the intimacy subscale
(alpha = .91) read 1) close to them, 2) attached to them, 3) bonded to them, 4) close-knit,
and 5) as a friend. Most widely used in the area of sports psychology research, positive
Although the above may not be an exhaustive list, it is clear that the diversity of
operational definitions and corresponding measurement tools suggest that the concept of
25
a relatedness need-supportive relationship needs clarification. There is, however, much
support for the notion that students’ perceived relatedness in the classroom has an
with students’ intrinsic motivation (Anderson, Manoogian, & Resnick, 1976; Ryan &
Grolnick, 1986) , perceived competence, self-esteem (Ryan & Grolnick, 1986), perceived
engagement in the classroom (Connell, 1990; Connell & Wellborn, 1991; Furrer &
Skinner, 2003; Klem & Connell, 2004; Ryan et al., 1994; Skinner et al., 2008).
firmly in attachment theory (Bowlby, 1973), both in terms of how the relationship is
conceptualized and how results are interpreted. Though, it might seem appropriate to
“student-teacher attachment,” this subheading was not used because other groupings that
are distinct enough in terms of current theory to be separated, also cite attachment theory
teacher’s point of view, the Student-Teacher Relationship Scale (STRS) was developed to
1994; Pianta & Nimetz, 1991; Pianta & Steinberg 1992). Its theoretical structure, derived
efforts to determine the correct empirical model for the factor structure of this scale have
26
produced mixed results. It has been found to consist of five factors: 1) conflict/anger,
which measures the struggle and emotional drain that the teacher experiences from the
which measures student willingness to share feelings and teachers attunement to the
students feelings, 4) dependency, which measures students’ unnecessary requests for help
from the teacher and overly strong reactions to separation from the teacher, and 5)
troubled feelings, which measures the student’s refusal of teacher support and teacher’s
worrying about the relationship (Pianta, 1994; Pianta & Steinberg, 1992). Yet, some
researchers prefer to stay closer to Bowlby’s (1973) model by using a three subscale
structure of the STRS including 1) closeness, 2) dependency, and 3) conflict (Birch &
Ladd, 1997, 1998). Within this framework the closeness subscale includes 11 items that
assess warmth and open communication. The dependency subscale is a 5 item index of
the degree to which the student is overly dependent on the teacher. The conflict subscale
consists of 12 items measuring friction or struggling between the student and teacher. It
should be noted that the dependency subscale has demonstrated relatively weak internal
consistency in past studies (e.g., alpha = .61-.69; Kesner, 2000; Palermo et al., 2007). A
15-item short version of the STRS has been used more recently (Baker, 2006; Crosnoe et
al., 2010; Rudasill, 2011), which consists of 8 items for closeness (alpha = .64-.86) and 7
items for conflict (alpha = .84-.91). Webb and Neuharth-Pritchett (2011) recently
examined the three-factor structure validity and measurement invariance of the full 28-
item STRS across 178 African American and 130 European American students. The full
sample was found to show poor model fit until 2 items were dropped. When the sample
27
was split by ethnicity, the pattern of factor loadings failed the test of configural
invariance and exploratory factor analysis revealed large discrepancies in factor loadings
for items on the closeness and dependency scales between groups. All this suggests that
the factor structure of the STRS is not clear; nonetheless, this scale is quite popular in the
assessment of scale properties have been reported in Greece (Gregoriadis & Tsigilis,
2008), Italy (Fraire et al., 2013), Netherlands (Koomen et al., 2012), Norway (Drugli &
combination with the typical teacher-report version of the STRS. Interestingly, the
maximum correlation between teacher- and child-report scores was only r = .30.
Henricsson and Rydell (2004) also found little (r = .21) to no relationship (r = .11)
Lisonbee, Mize, Payne, and Granger (2008) who, at best, found r = .20. This discrepancy
between teacher and student reports of the relationship is a clear example of issues related
constructs. Similarly, when Doumen et al. (2012) examined agreement between teacher
28
significant yet small correlation between observer and teacher ratings of closeness (r =
.31), conflict (r = .35), and dependency (r = .43). Still, the STRS in its various forms is
In a study on math achievement, Crosnoe et al. (2010) used the 15-item STRS to
measure closeness and conflict in the student-teacher relationship of 587 students in their
third- and fifth-grade years. They also measured math achievement using the Applied
Woodcock & Johnson, 1989) as well as basic vs. inferential instructional style using
observer ratings. Results suggested that both conflict and closeness predict increases in
math achievement from 3rd to 5th grade. This relationship is particularly strong for those
entering below average in Math and in classrooms that use inferential instruction.
Valiente et al. (2008) measured 264 students between the ages of 7 and 12-years old on
effortful control using the Early Adolescent Temperament Questionnaire (Capaldi &
Rothbart, 1992), teacher-student closeness and conflict using the STRS, as well as school
absences and GPA. The relationship variables were predictive of GPA and absences
beyond the effects of effortful control. In a cross cultural test of factor structure, Koomen
et al. (2012) delivered a translated version of the STRS to a Dutch sample of 2335
children ages 3-12. They also administered the Strengths and Difficulties Questionnaire
problems, and prosocial behaviors. The data suggested that the three-factor student-
29
the Dutch version of the STRS. All three factors were predictive of all five adjustment
achievement (Birch & Ladd, 1997; Crosnoe et al., 2010; Harme & Pianta, 2001; 2005;
Ladd & Burgess, 2001; Mantzicopoulos, 2005; O’Connor & McCartney, 2007; Pianta,
Belsky, Vandergrift, Houts, & Morrison, 2008; Pianta & Stulman, 2004), greater
classroom participation (Doumen et al., 2012; Ladd, Birch, & Buhs, 1999), less behavior,
internalizing, and attention problems along with higher work ethic, and frustration
tolerance (Pianta, 1994; Pianta & Nimetz, 1991; Pianta & Steinberg, 1992), greater social
competence with peers (Howes, 2000; Koomen et al., 2012; Pianta, 1994; Pianta &
Nimetz, 1991; Pianta & Steinberg, 1992), lower school avoidance, more cooperation, and
better self-discipline (Birch & Ladd, 1997), increased prosocial behavior and decreased
peer aggression (Birch & Ladd; 1998; Henricsson & Rydell, 2004; Koomen et al., 2012;
Palermo, et al. 2007; Stipek & Miles, 2008), less externalizing problems, and lower levels
of disruptive behavior (Baker, 2006; Henricsson & Rydell, 2004). Data also suggest
positive relationships act as a buffer against the negative effects of insecure maternal
mediate the relationship between effortful control and grades (Valiente, et al., 2008), and
they reduce the impact of various risk factors on academic achievement (Baker, 2006;
Harme & Pianta, 2005). Finally, in line with the contemporary interest in physiological
correlates of psychological phenomena, high scores on the conflict subscale of the STRS
30
have been shown to predict cortisol increases in children during teacher-child interactions
and conflict from students’ perspective is titled Young Children’s Appraisals of Teacher
measure, children are given a statement on a postcard and asked to place the card in a
mailbox if they agree and in a trashcan if they disagree, which generates a dichotomous
perceived opportunities for choice and variety in activities, and the conflict subscale has 8
Neuharth-Pritchett, 2003). Similar to what others have found when using both the STRS
and other student perception measures, the student reported Y-CATS data are modestly to
not-at-all correlated with teacher-reported STRS data. The subscales with the greatest
correlation are the Y-CATS warmth/closeness subscale and the STRS secure subscale,
yet these correlations range from only r = .25 in pre-school to r = .17 in first grade. Only
Sense of belonging. Initially considered for its relevance to school attrition rates
defined as the “extent to which students fell personally accepted, respected, included, and
31
supported by others in the school environment” (p. 80). Although broader in scope and
many similarities with them, which is why it has been included in this study.
values (Ryan et al., 1985). Further validating the inclusion of this construct with an
exploration of student-teacher relationship research, Demanet and Van Houtte (2012) use
seven items from the primary measure of school belongingness, discussed below, and
school misconduct. It is also common for researchers in this field, like Zumbrunn et al.,
research as though all are assumed to be measuring similar relationally supportive school
measure this relationship construct, demonstrates good internal consistency (α= .875) and
was a good predictor of effort and achievement in school (Goodenow, 1993). The PSSM
scale consists of 18 items which are averaged to create a single scale score. Items assess
perceptions such as the degree to which students feel able to be themselves, feel accepted,
32
feel their accomplishments are noticed by others, feel they are taken seriously, can talk
about problems with adults, feel included in activities, feel respected, and feel as though
they belong in the school. Goodenow’s (1993) belongingness scale has been adapted by
other educational researchers (e.g. Anderman, 1999; Anderman & Anderman, 1999;
Eccles & Midgley, 1989; Roeser et al., 1996), but the overall construct has not changed.
Variables associated with a high sense of belongingness include task valuing and
Goodenow & Grady, 1993), general school motivation (Goodenow & Grady, 1993), less
absenteeism (Nichols, 2008), reduced risky behaviors (Resnick et al., 1997), increased
positive affect (Anderman, 1999; Roeser et al., 1996), greater optimism, less depression,
less social rejection, and less behavioral problems (Anderman, 2002), greater academic
efficacy and intrinsic goal orientation (Freeman & Anderman, 2002; Zumbrunn et al.,
Anderman, 2002; Gutman & Midgley, 2000). Some additional terminology has been
used to refer to concepts nearly identical to the sense of belongingness. These include
identification with school (Finn, 1989), bonding with school (Kester, 1994), and
More recently, Kennedy and Tuckman (2013) used the PSSM scale along with
numerous other psychological and academic performance measures in their study of 671
college freshmen. Their data supported a model where perceived school belonging in the
eighth week of classes was positively related to performance goal orientation, self-
efficacy, mastery goal orientation, and negatively related to perceived stress (Kennedy &
Tuckman, 2013). Others have begun investigating the role that perceived school
33
belonging plays in students’ experience of negative affect and depressive symptoms
(Shochet & Smith, 2014; Shochet, et al., 2011). These studies involved 504 Australian
seventh and eighth grade students measured at three time points separated by twelve
months and then six months points using self-report measures for characteristics of the
classroom environment, students’ perceived school belongingness (the PSSM) and their
depressive symptoms (Shochet & Smith, 2014). School belonging was found to partially
symptoms, but both also showed unique predictive importance at each time point
The PSSM scale is clearly a reliable and valid measure of students’ perceived
school belonging. Across 26 studies reviewed by You, Ritchey, Furlong, Shochet, and
Boman (2011) the alpha coefficients ranged from .78 to .95. However, the factor
structure of the PSSM scale was brought into question (You et al., 2011) and it still does
not seem to be resolved (Ye & Wallace, 2014). Specifically, You et al. (2011) randomly
split their participants into two groups and identified three factors using exploratory
factor analysis with the first group (N = 256) and then validated the measurement model
using confirmatory factor analysis of the second group (N = 248). They named these
three factors caring relationships, acceptance, and rejection (You et al., 2011). Shochet et
al. (2011) reported further evidence supporting the importance of considering perceived
adolescent boys and rejection for predicting negative affect in adolescent girls. Demanet
and Van Houtte (2012) analyzed PSSM data from 11,872 high school aged students using
34
Varimax rotation in a Principle Components Analysis and found four component factors.
They determined that the first factor represented teacher related items and the other three
did not yield straightforward interpretation so they chose to accept a two factor
measurement model for the PSSM with one factor being teacher support and the other
being general school belonging, which yielded acceptable results in confirmatory factor
analysis (Demanet & Van Houtte, 2012). Most recently, Ye and Wallace (2014)
examined the factor structure of the PSSM scale in a diverse American sample of high
school students (N = 890) and found further evidence for a three dimensional factor
structure; however, the items in the second and third factors do not overlap completely
with the items in the You et al. (2011) three factor structure and to further complicate
things they named their three factors generalized connection to teachers, perception of
fitting in among peers, and identification and participation in school (Ye & Wallace,
2014).
Before moving beyond this section of the paper, which covers the fifth and final
exemplify again the primary problem that this dissertation is addressing. The Australian
State of Victoria’s Department of Education, Employment and Training (2000) use a very
similar term, school connectedness to represent a very different construct. Their school
connectedness scale consists of 4 items measuring the degree to which students look
forward to school, like school, enjoy schoolwork, and have fun learning at school (DEET,
Victoria, 2000). Using this conceptualization, Skues, Cunningham, and Pokharel (2005)
bullied in school. Though it may be interesting to note that students who are bullied
35
enjoy school less and do not look forward to going to school; it would be a mistake to say
conceptualized by Resnick et al. (1997) or Goodenow (1993). Skues et al. (2005) did not
directly make such a claim; rather, it serves as a great example of the broader issue being
addressed herein, which is the potentiality for misinterpretation when measures differing
greatly in content are used to measure constructs with similar, and sometimes the exact
same name. Kohl, Recchia, and Steffgen (2013) touch on this issue in their call for
stop creating new measures when multiple, well validated measures of the same
constructs already exist; again, echoing why measurement integration is so badly needed
and discussing student-teacher relationships and their importance for students’ social and
academic development. This review may not be fully comprehensive in scope; however,
it covers all the common perspectives and sheds light on the differences in measurement
tools and the similarities in empirical associations with positive social and cognitive
considerable, yet attempts toward such solutions are required in order to truly understand
student-teacher relationships and their implications for student growth and achievement.
With this purpose, two main questions need discussion: 1) what is being measured that
outcomes? and 2) what is the psychological mechanism responsible for these empirical
36
associations between student-teacher relationships and desirable social and academic
A general theme shared by multiple approaches is the notion that when measuring
Psychologists positing this construct often cite object relations theory (Behrends & Blatt,
1985; Fairbairn, 1952; Winnicott, 1965) or attachment theory (Bowlby, 1973) as the
intellectual background for their belief. Although their theoretical arguments may be
sound, the tools used to measure student’s relational schemas are broad, inconsistent, and
provide minimal direct evidence that such organized cognitive representations exist. For
1994), feelings and desires when with the teacher (Lynch and Cicchetti, 1992), feelings
while interacting with the teacher (Furrer & Skinner, 2003), a combination of behaviors
involving the teacher, feelings when around the teacher, and desire to emulate the teacher
(Ryan, Stiller, & Lynch, 1994); or, perhaps furthest from directly measuring students’
relational schemas, some measures ask teachers to report their feelings about the
relationship and their perceptions of student behavior (Pianta & Steinberg 1992).
So what are these mysterious relational schemas that are measured at various
degrees of directness by the aforementioned instruments? One could possibly argue that
37
questionnaire items such as these. This seems unlikely, due to the association between
however, without direct empirical evidence regarding the structure and function of
needed.
dynamic, organized schemas derived from, and used on-the-fly for socializing
interactions with others (Ryan, Avery, & Grolnick, 1985). From psychodynamic to social
cognitive perspectives, this dynamic set of schemas, which contain both knowledge and
affective information (Fiske, 1982), go by many different names. Some of these names
include internal working models of self in relation to others (Bowlby, 1973; Bretherton,
1985; Crittenden, 1990), relational models (Fiske, 1992), relational self (Andersen &
Chen, 2002), self-with-other unit (Ogilvie & Ashmore, 1991), interpersonal schemas
1989), and relational schemas (Baldwin, 1992; Planalp, 1987). Because this construct
exists by so many names, to reduce confusion the term relational schema has been and
there would be some direct evidence for the existence of this construct. In particular,
methods used in research on social cognition seem well suited for investigating how
relationships are perceived, stored, processed, and recalled. Disappointingly, until the
mid-1990’s, social cognitive research has focused on the aspects of social perception
38
considered separately. For instance, a researcher might study how people perceive others
(e.g., Cantor & Mischel, 1979; Fiske & Cox, 1979; Schneider, 1973), how people
perceive themselves (e.g., Epstein, 1973; Markus, 1977; Rogers, 1977), or how people
construe situations (e.g., Cantor, Mischel, and Schwartz, 1982; Schank & Abelson, 1977).
However, some empirical support for the interconnected organization of these cognitive
structures has been established (Baldwin, 1992). Although the content, structure, and
function of relational schemas have never been directly investigated in the context of
perceiving the teacher as a powerful, rejecting, criticizer. This student would likely
maintain if-then scripts such as “if I make a mistake then the teacher will scold me.”
relational schemas should bring about social cognitive phenomena such as the drawing of
attention to schema relevant information, the filling in of information gaps, and biasing of
and the process of spreading activation as defining aspects of organized schemas (Higgins
& Bargh, 1987), it is safe to say that Baldwin and his colleagues have illustrated the
existence of relational schemas and some of their perceptual influences. For example,
Baldwin and Holmes (1987) primed female participants’ relationships with their parents
and other participants with their campus friend relationships; then, in an ostensibly
unrelated task the researchers had participants rate their enjoyment of various stories. As
39
expected, participants with a primed parent-relational schema rated a sexually permissive
story as less enjoyable and less exciting than the other participants. In a second study,
difficult task. Participants primed with a shallow relationship were more likely to feel
badly and attribute their failure to “something about me” (Baldwin & Holmes, 1987).
These results have been replicated using the simple priming technique of exposure to the
To further explicate the underlying if-then script activation responsible for these
findings, Baldwin and Sinclair (1996) demonstrated that participants primed with a
a lexical decision task. This type of if-then script activation is an important mechanism
for the influence of relational schemas on thought (see Baldwin, 1997). Again using the
exists in support of the relational schema as a social cognitive construct. Although direct
evidence has not yet been collected, it seems plausible to believe that such schemas also
40
Fitting relational schemas with previous student-teacher relationship
measuring aspects of relational schemas. The major aspects of a relational schema of this
kind would consist of the student’s perception of self, perception of the teacher, and
associated interpersonal scripts. Taking a closer look at the measure of school belonging
relational schema for their relationship with others at school. As such, it seems to have 6
items that tap into students’ self-perceptions, such as “sometimes I feel as if I don’t
belong here,” 4 items examining students’ perception of other, such as “most teachers at
(name of school) are interested in me,” and 8 items which relate to interpersonal scripts,
such as “people here notice when I am good at something.” Of the 6 sample items
school community, 4 relate to students’ interpersonal scripts, for instance “the teacher in
my class asks the students to help decide what the class should do,” and 2 tap perceptions
of others in the classroom, such as “the students in this class really care about one
another.”
As for the self-determination theorists, Ryan et al. (1994) specifically state that
they intend to measure relationship schemas and Lynch and Cichetti (1992; 1997) claim
when explaining their relatedness measure as do Skinner and her colleagues (Furrer &
Skinner, 2003; Skinner & Belmont, 1993). Accordingly, the IAA (Greenberg, 1982 as
41
teacher such as trustworthiness and understanding as well as items examining students’
interpersonal scripts related to sharing feelings and accomplishments with the teacher.
Ryan et al.’s (1994) school utilization and teacher emulation components reflect the same
perception of teacher and interpersonal script foci. Skinner and Belmont’s (1993)
appreciating, understanding, helpful) and their interpersonal scripts (e.g. if I need the
teacher then they will be available for me); whereas, the Furrer and Skinner (2003)
relatedness measure only taps affective aspects of the students’ if-then scripts, which is
also true for the Laguardia et al. (2000) measure. The Wellborn and Connell assessment
tool used by Lynch and Cichetti (1992; 1997) focuses on the measurement of both
The measure of pedagogical caring used by Wentzel (1994), only taps students’
perceptions of the teacher, specifically, whether the teacher cares about a number of
things related to the student’s school life. Feldlaufer et al.’s (1988) measure of teacher
supportiveness contains a few complex items that could arguably fit into many different
parts of the model. First, they appear to tap students’ perceptions of the teacher as caring
and fair; however, many of the items would require activation of students’ if-then
interpersonal scripts to generate a response because the items ask about specific teacher
behaviors in relation to specific student behavior. In the end, the focus on specific
teacher behaviors in these items makes them sound most like teacher involvement or
teacher use of punishment, which are predictors of the relationship in the model. For
example, the item worded “this teacher threatened to give me more work, tests, or to
lower my grade to control my behavior” could imply that the teacher is uncaring or
42
unfair, but the if-then script about misbehavior leading to threats would be activated and
the item would fit well as a measure of the teacher’s use of punishment.
The STRS (Pianta, 1994) does not directly measure any aspect of students’
some of their beliefs about students’ perceptions. Specifically, teachers are asked to
with the student (e.g. if upset, this child will seek comfort from me). Mantzicopoulos and
Neuharth-Pritchett’s (2003) Y-CATS measure on the other hand, does tap students’
relational schemas by asking young children to indicate their perception of the accuracy
behaviors, which would require the activation of the students’ if-then interpersonal
scripts.
relationship measurement strategies through the social cognitive, relational schema lens.
Two important questions next emerge: 1) what are the most basic characteristics of a
challenge of this study would be to tease apart the effect sizes for each outcome measure
for each factor measured within each student-teacher relationship conceptualization. This
43
process could become inherently confounded due to the various questionnaire items and
Another way to approach the question would be to perform a large study wherein
all the current major perspective on defining high-quality student teacher relationships
relationship characteristics. Perhaps some new characteristics would emerge from deeper
reflection on the relational schema concept as well. For instance, current perspectives do
not tend to measure much about students’ perceptions of self. The items from all
relationships and which items best measure such factors. In a subsequent study, these
factors could be examined for predictive power in relation to each of the outcome
Answers to the second question require additional research as well; however some further
Research has provided much evidence that high quality student teacher
relationships are associated with pro-social classroom behaviors and social competence
academic achievement and emotional adjustment (e.g., Birch & Ladd, 1996; 1997;
Goodenow, 1993; Ryan et al., 1994). Unfortunately, exactly how students’ perceptions
of high quality relationships with their teachers lead to the desirable outcomes is not
entirely known. Actually, there are multiple ways to explain these social and cognitive
44
developmental effects of student-teacher relationships, but the critical research has not yet
One explanation comes directly from attachment theory (Bowlby, 1973). This
theory would suggest that high quality student-teacher relationships provide students with
a sense of security allowing their exploration of the environment, which leads to the
influence the construal of behavior (Bargh & Pietromonaco, 1982; Higgins, Rholes, &
Jones, 1977; Srull & Wyer, 1979). Correspondingly, the activation of a particular
suggested that perhaps teachers simply like high achieving and socially developed
students more, or at least find them easier to befriend. This would imply that positive
relationships might actually be the outcome and not the cause. Of course, something
along those lines could be true in addition to the relationship having an impact on further
which data suggest some support for (Birch & Ladd, 1998; Skinner & Belmont, 1993;
Yet another possibility is that the correlational findings are due to a kind of third
45
simultaneously generate greater social and cognitive developmental outcomes. There is
some support for this notion in the relationship between instructional practices and
(1985). Using a similar line of thought, perhaps high quality relationships develop
between students and teachers when the student’s behavior is moderated by other
cognitive development.
It is likely that all of these explanations have some merit. It is also likely that
each explanation has more or less merit depending on the outcome under consideration.
However, there is another explanation which has powerful implications for its potential
internalization processes, which provide students with motivation to behave well and
achieve academically.
facilitate students’ internalization of socially prescribed beliefs, values, and goals (Ryan,
1993; 1995). Simply put, if students feel that their teacher truly understands and cares
about them, then they are likely to accept the teacher’s social and task-related goals as
their own and willfully pursue them. Through the internalization process, goals that
originate externally (from the teacher) become valued or deemed important by the student
Without internalization, goals will only be pursued for purely extrinsic reasons such as
46
pursuit may be energized to avoid feelings of guilt or to preserve one’s sense of self-
worth (introjection). Although the levels of internalization are labeled for ease of
socializing agents (Hodgins, Koestner, Duncan, 1996; Ryan, 1993, 1995; Ryan &
quality is defined by the degree to which the relationship fulfills the basic psychological
connection and belongingness with others. As was also revealed earlier, the best way to
clear. What is clear is that measuring from the student’s perspective is crucial if the
effects of relatedness on facilitating internalization processes within the student are the
focus. This is not to say that the teacher’s perception of the relationship (e.g. STRS,
Pianta & Steinberg, 1992) is useless; actually, there are many cases when the teacher’s
process, then measuring the relationship from the student’s point of view makes the most
sense.
In fact, Sandler and Rosenblatt (1962), extended the Freudian perspective of learning as
47
representation on the basis of another (usually an object) representation as a
(1968) described internalization as “all those processes by which the subject transforms
real or imagined regulatory interactions with his environment, and real or imagined
characteristics of his environment, into inner regulations and characteristics” (p. 9). Also
taking a Freudian perspective was Fleming’s (1972) discussion of the “learning alliance”
as being the teacher-student form of the “therapeutic alliance” between psychoanalyst and
many other popular psychological theories (e.g., Erikson, 1950; Fromm, 1956; Loevinger,
Some of the other research perspectives reviewed earlier in this paper also
propose the internalization explanation as well. Specifically, the Grussec and Goodenow
of organismic integration theory (Deci & Ryan, 1985; Ryan, 1993), as do the
The relationship between a student and teacher is, at the very least, a strong
predictor of many desirable social and cognitive developmental outcomes. It likely plays
a causal role in producing such outcomes, though the mechanisms involved need further
educational and psychological researchers. Although the tools used for measurement
vary greatly in terms of both format and content, it is interesting how similar many of the
48
study outcomes are. There is need for a theory-based clarification of the relationship
construct responsible for such outcome consistency. Social cognitive psychologists have
provided strong empirical support for relational schema theory, which appears to hold
relationship effects as well. Further, relational schema theory might be a good place to
start with an attempt to measure more directly and thoroughly the underlying relationship
construct in students’ psychological worlds, which ultimately impact their social and
development is also not entirely understood and many explanations have been proposed.
These too need further study. However, the first step toward this end is to improve our
49
CHAPTER III. METHODS
Purpose
consistent in educational research. The purpose of this study is to examine the many
explore an integrative theory and measurement approach that accounts for their
perceptions was gathered using instruments from all the perspectives simultaneously.
Participants
via email for participation in this research with the promise of entry into a drawing for
$50 as an incentive for participation. The survey was opened by 893 participants. Of
those who started the survey, 238 participants were deleted for not officially submitting
the survey. These 238 participants had submitted demographic data and received the
complete any of the subsequent 179 survey items referencing that relationship. Next, 4
more participants were deleted because they responded “no” to the demographic question
about being 18 or older, which immediately submitted their survey even before they
received the 179 student-teacher relationship items. Of the remaining 651 participants,
266 provided partially incomplete data. A visual scan of the raw data clearly revealed
50
that while some participants had large amounts of the survey items unanswered, the vast
majority of these 266 participants had only a few missing item responses. The criteria of
95% complete data or greater was used to retain participants, which is in line with the 5%
missing data cut-off suggested by Shafer (1999). It also results in a data set that retains
Schlomer, Bauman, & Card (2010). In other words, all participants who had missed 9 or
more of the 179 student-teacher relationship items were deleted. This removed 19 more
participants from the sample. Finally, 4 more participants were removed for not
responding to key demographic grouping questions such as gender and years since high
school graduation, which left 628 participants in the sample. Because the questionnaire
included 179 student-teacher relationship questions, the remaining 628 participant’s data
standard deviation of responses was .60, therefore all 628 of these participants were
retained.
Of the 628 that remained in the sample, 136 were male and 492 were female. The
participants were randomly assigned to report their most recent experiences with a high
school math (N = 207), English (N = 202), or science (N = 219) teacher. Because these
high school, it is important to report how long it had been since they experienced that
relationship. Accordingly, 152 participants experienced the relationship less than one
year ago; whereas 150, 103, 88, 37, and 98 participants experienced the relationship two,
three, four, five, and more than five years ago respectively. Participants reported on
whether the teacher’s gender was male (N = 322) or female (N = 306) as well.
51
Instrumentation
been given the title Composite Student-Teacher Relationship Instrument (C-STRI). This
instrument combines twelve instruments that have been used to measure the relationship
between students and their teachers, as well as two instruments used to measure a slightly
broader relationship between students and their classroom or school. To review, Table 1
contains a descriptive list of these fourteen instruments. After considering the risks of
creating too lengthy an instrument, all items from these scales were combined into a
single composite instrument for use in this study. There were two primary reasons for
this: 1) it allows for individual scale scores, predictive validity, and internal consistency
the potential for items with strong empirical importance, but less face validity, to be
incidentally removed in the process of reducing the length of the composite instrument.
cohesive, composite instrument to allow all items from all fourteen instruments to be
administered as one instrument with random item ordering and minimized participant
confusion. All items from each of the 14 instruments, kept grouped by instrument of
origin, with original wording, and compiled into a single document for a total of 170
The original instrument response scales varied from 4-point to 7-point Likert style
formats. The response scales were thus removed and replaced with a 6-point scale to
standardize potential response variance in the C-STRI. Possible responses ranged from
strongly disagree to strongly agree. Although the question of an ideal number of scale
52
items remains unanswered despite decades of research (Preston & Colman, 2000), a 6-
point scale was selected in accordance with Miller’s (1956) discussion about the limits of
Next, item statements were slightly modified only as necessary to 1) ask for a
perspective, 2) refer to the relationship in the past tense as opposed to present tense, and
general or the whole classroom. The new list of 170, conceptually cohesive items, still
perspective, item 114 (Appendix A) was originally worded “this student trusts me” in the
STRS (Pianta & Nimetz, 1991) and was reworded to “I trusted this teacher” for use as
item 114 in the C-STRI (Appendix B). As an example of referring to the past instead of
present-tense, item 27 (Appendix A) was originally worded “my teacher really cares
about me” in the TASC (Belmont et al., 1988) and was reworded to “this teacher really
cared about me” for use as item 27 in the C-STRI (Appendix B). As an example of
changing the wording to refer to a specific teacher rather than teachers in general (and
present to past-tense), item 65 (Appendix A) was originally worded “teachers here are not
interested in people like me” in the PSSM (Goodenow, 1993) and was reworded to “this
teacher was not interested in people like me” for use as item 65 of the C-STRI.
53
added to assess the comparative predictive validity of each original scale and to
strengthen the measurement of those outcomes in the hypothesized structural model. The
full set of C-STRI items, worded as they were when administered to participants, is
54
Y2
Y5
Y59
Y66
Y71
Y90
Y91
Y93 ή1
Predictor:
Y94
Academic Support
Y102
Y103
Y105
Y147
Y151
Y152
Y156
Y47
Y48
Y50
Y51
Y52
Y53
ή2
Y104 Predictor:
Y106 Punishment
Y107
Y108
Y111
Y112
Y120
Y122
Y149
55
Y7
Y24
Y32
Y33
Y34
Y35
Y36
Y37
Y38
ή3
Y39 Predictor:
Involvement
Y56
Y60
Y67
Y95
Y96
Y97
Y128
Y129
Y150
Y154
Figure 1. Continued
56
Y98
Y99
Y100
Y101
Y109
Y110
ή4
Y113
Predictor:
Y157 Autonomy
Support
Y158
Y159
Y160
Y161
Y162
Y163
Y164
Y165
Y166
Figure 1. Continued
57
Y1
Y4
Y6
Y26
Y27
Y28
Y49
Y61
Y64
Y65
Y70
Y74
Y75
Y76 ή5
Caring
Y77
Y78
Y79
Y80
Y81
Y82
Y83
Y84
Y87
Y88 ξ1
Perceptions of
Y89
Teacher
Y125
Y148
Y153
Y8
Y13
Y15
Y22 ή6
Y25 Awareness/
Y29
Understanding
Y30
Y31
Y115
Figure 1. Continued
58
Y10
Y54
Y55
Y57
Y68
Y69
Y85
Y86
Y114 ή7
Y116
Closeness
Y124
Y130
Y131
Y132
Y133
Y134 ξ2
Perceptions of Self
Y135
Y136
Y137
Y138
Y139
Y21
ή8
Y23 Affective
Y127 Experience
Y167
Y168
Y169
Y170
Figure 1. Continued
59
Y9
Y14
Y16
Y17
ή9
Y18 Social Support
Y19
Seeking
Y20
Y63
Y119
Y142
Y155 ξ3
Interpersonal Scripts
Y12
Y40
Y41
Y42
Y43
ή10
Y44 Emotional
Y45
Reliance
Y46
Y117
Y126
Figure 1. Continued
60
Y73
Y143
Y144
Y145 ή11
Y146
Outcome
Internalization
Y171
Y172
Y173
Y174
Y3
Y11
Y58
Y62
Y72
Y92
Y118
ή12
Y121
Outcome
Y123 Engagement
2
Y140
Y141
Y175
Y176
Y177
Y178
Y179
ή13
Outcome
Y180 Academic
Achievement
Figure 1. Continued
61
To deliver the full instrument, all 180 C-STRI items were entered into Qualtrics,
an electronic survey delivery software program. Using this software, items were
presented to each participant in random order to avoid order effects like item familiarity,
which might result from the use of numerous similar items, and item fatigue, which might
result from the extraordinary length of this instrument. Using Qualtrics also allowed the
C-STRI to be administered online to simplify data collection and eliminate human errors
perceiving self, others, and situations to offer a cohesive, social cognitive theory of
perception, and interpersonal scripts, which provided a foundation for the three latent
62
ή6
ή5
Awareness/
Caring
Understanding
ή1 ή9 ή10
Academic Social Support Emotional
Support Seeking Reliance
ή11
ξ1 Internalization
ή2 Perceptions of
Punishment Teacher
ξ3
ή12
Interpersonal
Engagement
Scripts
ή3 ξ2
Involvement Perceptions of
ή13
Self Academic
Achievement
ή4
Autonomy
Support
ή8
ή7
Affective
Closeness
Experience
Baldwin’s theoretical relational schema structure also provided guidance for the a
priori coding categories for performing a content analysis (Stemler, 2001) of the C-STRI
item statements to determine which items would likely represent each latent variable in
the measurement model (Figure 1). In other words, all original scale items were sorted
into groups that represented perceptions of self, perceptions of the teacher, and if-then
type interpersonal script statements. Emergent coding was used on the items that did not
fit into the a priori categories. Those items fit into the two primary emergent groupings
of teacher behaviors and student behaviors. Teacher behaviors are external stimuli that
63
whereas student behaviors are outcomes influenced by students’ interpersonal scripts.
Original scale items that reflected statements about how students perceived their teacher
fell into the two a priori factor groupings labeled caring and awareness/understanding.
Original scale items that reflected statements about students’ self-perception fell into two
a priori factor groupings as well, which were labeled closeness and affective experience.
Original scale items that reflected if-then statements about how students’ expected they
or their teacher might behave (interpersonal scripts) fell into the two a priori factor
groupings labeled emotional reliance and social support seeking. The specific items
Quasi-Independent Variables
gender, teacher gender, and the subject matter taught by the teacher of reference in
the hypothesized model. Participants’ high school graduation year and the year in high
school that they took the class was also recorded and combined to determine how far
back the participants were being asked to remember and examine consistency between
participants remembering relationships less than two years ago and participants
remembering four or more years ago. Participant gender, teacher gender, high school
graduation year, and year they took the class in high school were recorded by asking the
participants to report it at the start of the survey. Subject matter taught by the teacher of
three versions of the C-STRI were used. The only difference between the three versions
64
was the wording of a single sentence in the instructional paragraph, which preceded the
survey items. This instruction asked students to “use memories of experiences with their
most recent high school _______ teacher” while completing the survey questions. The
blank was filled in with math, science, or English in versions one, two, and three
respectively. These class topics were selected because they are considered common core
classes that all students would have typically taken during junior and senior year in high
school.
Numerous items from the original instruments were categorized into four factor
students’ relational schemas; however, they are teacher behaviors and not part of the
relational schema. As outlined in Appendix C, they are grouped as factors that represent
things that teachers do, they are not represented as parts of the students’ internal
relationship representation per se. Instead, they are treated as predictor variables in the
hypothesized structural model (Figure 2). These four quasi-independent variables are
and encouraging student academic achievement, providing support when the student did
not understand something, and assigning work at the appropriate level for the student.
teaching behaviors such as threatening the student, using sarcasm, imposing unrealistic
expectations, criticizing the student, and being mean. Essentially, it represents teacher
behaviors that create conflict between the student and the teacher.
65
Involvement. Items in this grouping relate to things teachers do to show students
that they appreciate them, understand them, and respect them. Examples of these
behaviors include remembering the student’s birthday, being available for the student in
times of need, and smiling at the student. These teacher behaviors show students that
teachers see them as individuals whose feelings are important. While recognizing that
these behaviors are central to relationship development, they are still teacher behaviors
and thus not included in the model as part of the students’ relationship representation.
in establishing classroom rules, and using student input/feedback, and providing rational
for assignments. The opposite of these strategies would be strategies that impose a
controlling atmosphere.
Caring. This factor represents the greatest number of items from the original
scales. These 28 items relate to the student’s perceptions of whether the teacher was
warm or friendly toward them, cared about them, and generally liked them. The large
number of items that categorically fit into this construct from numerous original
statements of the student’s perception of how much the teacher knew about them and
66
Closeness. This factor also comprised a very large number of items. The 21
items in this grouping represent perceptions of the student’s trust in the teacher and their
Social Support Seeking. The 11 items included in this factor represent how
willing the student would be to seek out help from the teacher if experiencing a problem.
Emotional Reliance. These 10 items are quite similar to social support seeking
items with the difference that all items in this factor relate specifically to a willingness of
the student to turn to the teacher for support when experiencing strong negative emotions
such as fear, anxiety, and frustration, or strong positive emotions like pride and
excitement.
Outcome Variables
categorized into two student outcome factor groupings, internalization and engagement,
which represent characteristics of the student’s relationship with his/her schoolwork and
not the student’s relationship with the teacher. One more outcome, not measured by any
of the original tools, but included in the study for the purpose of examining predictive
Internalization. Items included in this grouping from the original scales represent
the students’ desire to be like or emulate the instructor. Because the original scales were
items were added to potentially strengthen the measurement of this outcome, which
67
represents students’ internalization of the purpose, importance, or value in learning the
subject matter. These four items are listed under the internalization sub-heading in
Engagement. Eleven items in this category were taken from the original scales
that represent both the emotional experience of being in class and participatory behaviors.
Again, because the original scales were not intentionally constructed to represent
outcomes such as engagement, five additional items were added to represent other
aspects of cognitive, behavioral, and affective engagement in the class for a total of 16
items. The five items that are not from the original fourteen scales are listed under the
report the grade that they earned in the class taught by the teacher of reference. Whereas,
all other C-STRI items will be responded to using a 6-point scale ranging from strongly
agree to strongly disagree, for this item the response options will be A, A-, B+, B, B-, C+,
Procedure
they would forward an email invitation to participate in this project. At one university,
the other university, all instructors teaching 100 or 200 level courses in the Psychology,
68
forward on to their students. Entry into a drawing for $50 was used as an incentive to
participants was desired (200 participants per course subject condition), approval to
mass-email the survey to the entire undergraduate student body at one of the universities
was pursued and granted. Following this distribution, the participant number grew to 580
responses. The mass-email of this survey was sent out a second time as a reminder,
which brought the final number of responses to 893 responses. The process of
winnowing the 893 responses down to the 628 participant included in the analyses is
Survey administration. The survey link, which was embedded in the email
informed consent page (appendix E). Interested students read the informed consent page
and indicated their consent to participate by entering their name and email address into
the consent page and clicking the submit button. After the participant submitted consent
questions included the year the participant graduated high school, the participant’s
gender, and if the participant was 18 years old or older. If the participant responded no to
the question about being 18 years old or older, Qualtrics skipped directly to the end of
survey thank you note (Appendix G). For all participants who answered yes to the 18
years or older question, Qualtrics randomly assigned them to one of the three course
STRI introduction statement (Appendix H). After reading the introduction statement,
69
participants clicked the “continue” link and Qualtrics administered all of the remaining
C-STRI questions. The first three questions directly following the introduction statement
asked participants the year of high school that the class took place, the letter grade that
the participant earned in the class, and the gender of the teacher for the class. The next
179 items were presented in a randomized order for each participant. See Appendix I for
Hypotheses
items from each of the original instruments will demonstrate acceptable internal
tools will demonstrate acceptable predictive validity in the expected direction as related
Hypothesis 3. Items from the all the measurement tools taken together will fit
demonstrate good fit. Specifically, the four variables (academic support, punishment,
involvement, and autonomy support) will predict the three outcome variables
schema. The schema will be represented by three latent variables 1) perceptions of self
with the two observed variables closeness and affective experience, 2) perceptions of
teacher with the two observed variables caring and awareness/understanding, and 3)
70
interpersonal scripts with the two observed variables social support seeking and
emotional reliance.
Data Analysis
consistency and fit of the data to the proposed factor structure of each original scale and
respective subscales.
each original scale in relation to the outcome variables internalization, engagement, and
academic achievement.
consistency and fit of the data to the hypothesized factor structure proposed in Figure 1.
First, the measurement model of predictor variables was examined. Second, the
measurement model of outcome factors was examined. In cases where the observed data
did not reasonably fit the hypothesized model, model modifications such as item
elimination and allowing correlated error variance were utilized to improve the goodness
of fit. Minimum acceptance of reasonable fit required the model to have an RMSEA
value of .08 or less (Browne & Cudeck, 1993) and a CFI value of .90 or greater (Hu &
Bentler, 1999).
fit of the hypothesized full structural model of latent variable relationships as displayed in
71
Figure 2. Again, the model underwent modifications until the acceptable fit was reached
where the minimum acceptance of reasonable fit required the model to have an RMSEA
value of .08 or less (Browne & Cudeck, 1993) and a CFI value of .90 or greater (Hu &
Bentler, 1999).
multicollinearity problems among its factors, exploratory factor analysis was utilized to
reveal the structure of the data from a purely empirical starting point. The principal
components analysis method of extraction was used to reduce the set of items down to a
small number of factors that could potentially be used to create composite scores for use
VARIMAX rotation, developed by Kaiser (1958), was used. The number of factors to
For decisions of item elimination, the following criteria were used. Items were
eliminated that did not have a primary factor loading of .6 or greater. Items were also
eliminated if the gap between primary and cross-loadings was less than .2. Items with
cross-loadings of .3 or above were eliminated as well. Finally, after the latent factors
were identified, the remaining items were examined for meaningful and useful
membership to their factor by reading over the item wording to assess face validity and to
determine the extent to which the item appeared to be redundant with other items.
72
CHAPTER IV. RESULTS
Descriptive Statistics
The average completion time for all 628 participants was 24.02 minutes.
However there were a few participants who must have begun/opened the questionnaire,
left it to do something else, and went back to complete/submit it later. In a few cases
several hours passed between the questionnaire’s start and finish time. Due to the
outliers, it is important to note that the median of all 628 completion times was 16.70
minutes. Of the 610 participants who completed the questionnaire in less than one hour,
the average completion time was 17.93 minutes. C-STRI item codes, item wording,
whether an item was reverse scored or not, and the item descriptive statistics for all items
in the C-STRI that utilized the 6-point scale and were presented to participants in random
73
Table 2. Continued
This teacher helped me to talk about my
A17_IPPA_C 628 1.0 6.0 3.39 1.41
difficulties.
I could count on this teacher when I needed to get
A18_IPPA_C 628 1.0 6.0 3.41 1.43
something off my chest.
If this teacher knew something was bothering me,
A19_IPPA_C 628 1.0 6.0 3.60 1.37
he/she asked me about it.
Talking over my problems with this teacher made
A20_IPPA_A X 628 1.0 6.0 4.61 1.30
me feel ashamed or foolish.
A21_IPPA_A I got upset easily around this teacher. X 628 1.0 6.0 5.00 1.20
A22_IPPA_A I get upset a lot more than this teacher knew about. X 628 1.0 6.0 3.70 1.52
A23_IPPA_A I felt angry with this teacher. 628 1.0 6.0 4.94 1.29
A24_IPPA_A I didn’t get much attention from this teacher. X 628 1.0 6.0 4.33 1.28
This teacher didn’t understand what I was going
A25_IPPA_A X 628 1.0 6.0 4.03 1.42
through in those days.
A26_Inv_Aff This teacher liked me. 628 1.0 6.0 4.63 1.18
A27_Inv_Aff This teacher really cared about me. 628 1.0 6.0 4.22 1.32
This teacher didn’t seem to enjoy having me in
A28_Inv_Aff X 628 1.0 6.0 4.94 1.21
his/her class.
A29_Inv_Att This teacher knew a lot about me. 628 1.0 6.0 3.56 1.50
A30_Inv_Att This teacher knew me well. 628 1.0 6.0 3.75 1.46
A31_Inv_Att This teacher didn’t understand me. X 628 1.0 6.0 4.45 1.36
A32_Inv_DR This teacher spent time with me. 628 1.0 6.0 3.74 1.34
A33_Inv_DR This teacher talked with me. 628 1.0 6.0 4.54 1.23
A34_Inv_Dep This teacher was always there for me. 628 1.0 6.0 3.94 1.33
A35_Inv_Dep I could count on this teacher to be there for me. 628 1.0 6.0 4.10 1.38
I could rely on this teacher to be there when I
A36_Inv_Dep 628 1.0 6.0 4.02 1.39
needed him/her.
A37_Inv_Dep This teacher was never there for me. X 628 1.0 6.0 4.79 1.26
I couldn’t depend on this teacher for important
A38_Inv_Dep X 628 1.0 6.0 4.46 1.40
things.
I couldn’t count on this teacher when I needed
A39_Inv_Dep X 628 1.0 6.0 4.43 1.44
him/her.
If I were feeling alone or depressed, I would have
A40_ERQ 628 1.0 6.0 3.42 1.56
been willing to turn to this teacher.
If I were feeling anxious or scared about
A41_ERQ something, I would have been willing to turn to this 628 1.0 6.0 3.60 1.52
teacher.
If I were feeling excited about something
A42_ERQ happening in my life, I would have been willing to 628 1.0 6.0 3.81 1.49
turn to this teacher.
If I were feeling very bad about myself and needed
A43_ERQ a boost, I would have been willing to turn to this 628 1.0 6.0 3.53 1.56
teacher.
If I were feeling overwhelmed, I would have been
A44_ERQ 628 1.0 6.0 3.87 1.52
willing to turn to this teacher.
If I were feeling frustrated or angry, I would have
A45_ERQ 628 1.0 6.0 3.65 1.50
been willing to turn to this teacher.
If I were feeling proud of my accomplishments, I
A46_ERQ 628 1.0 6.0 4.07 1.46
would have been willing to turn to this teacher.
This teacher criticized me for turning work in late
A47_Friend X 628 1.0 6.0 4.19 1.45
or failing to turn in assignments.
This teacher told me publicly that I might get a bad
A48_Friend X 628 1.0 6.0 5.22 1.20
grade in that subject.
A49_Friend This teacher was warm and supportive. 628 1.0 6.0 4.31 1.38
74
Table 2. Continued
This teacher seemed pessimistic about my ability to
A50_Friend be self-disciplining and responsible for my own X 628 1.0 6.0 4.79 1.31
behavior.
A51_Friend This teacher expected me to make stupid mistakes. X 628 1.0 6.0 4.98 1.19
A52_Friend This teacher used sarcasm. X 628 1.0 6.0 2.87 1.44
This teacher threatened to give me more work,
A53_Friend X 628 1.0 6.0 5.44 0.92
tests, or to lower my grade to control my behavior.
A54_Rel When I was with this teacher I felt accepted. 628 1.0 6.0 4.39 1.30
When I was with this teacher I felt like someone
A55_Rel 628 1.0 6.0 3.39 1.42
special.
A56_Rel When I was with this teacher I felt ignored. X 628 1.0 6.0 4.78 1.28
A57_Rel When I was with this teacher I felt unimportant. X 628 1.0 6.0 4.76 1.32
A58_PSSM I felt like a real part of this teacher’s class. 628 1.0 6.0 4.39 1.36
This teacher noticed when I was good at
A59_PSSM 628 1.0 6.0 4.53 1.25
something.
A60_PSSM This teacher took my opinions seriously. 628 1.0 6.0 4.39 1.27
A61_PSSM This teacher was interested in me. 628 1.0 6.0 4.04 1.31
Sometimes I felt as if I didn’t belong in this
A62_PSSM X 628 1.0 6.0 4.47 1.48
teacher’s class.
A63_PSSM I could talk to this teacher if I had a problem. 628 1.0 6.0 4.08 1.42
A64_PSSM This teacher was friendly to me. 628 1.0 6.0 4.88 1.17
A65_PSSM This teacher was not interested in people like me. X 628 1.0 6.0 4.79 1.33
This teacher made sure I was included in lots of
A66_PSSM 628 1.0 6.0 4.26 1.27
activities in class.
A67_PSSM I was treated with respect by this teacher. 628 1.0 6.0 4.89 1.21
A68_PSSM I felt very different from this teacher. X 628 1.0 6.0 3.80 1.45
A69_PSSM I could really be myself around this teacher. 628 1.0 6.0 4.29 1.40
A70_PSSM This teacher respected me. 628 1.0 6.0 4.78 1.22
A71_PSSM This teacher knew I could do good work. 628 1.0 6.0 5.02 1.01
A72_PSSM I wished I was in a different teacher’s class. X 628 1.0 6.0 4.65 1.66
A73_PSSM I feel proud of knowing this teacher. 628 1.0 6.0 4.39 1.49
A74_PSSM This teacher liked me the way I was. 628 1.0 6.0 4.54 1.20
It was hard for people like me to be accepted by
A75_PSSM X 628 1.0 6.0 4.78 1.32
this teacher.
A76_CLM_SS This teacher really cared about me. 628 1.0 6.0 4.19 1.34
This teacher thought it was important to be my
A77_CLM_SS 628 1.0 6.0 3.32 1.36
friend.
A78_CLM_SS This teacher liked me as much as other students. 628 1.0 6.0 4.52 1.25
A79_CLM_SS This teacher cared about my feelings. 628 1.0 6.0 4.24 1.27
A80_CLM_AS This teacher cared about how much I learned. 628 1.0 6.0 4.82 1.20
A81_CLM_AS This teacher liked to see my work. 628 1.0 6.0 4.51 1.16
A82_CLM_AS This teacher liked to help me learn. 628 1.0 6.0 4.74 1.24
This teacher wanted me to do my best in
A83_CLM_AS 628 1.0 6.0 5.15 0.99
schoolwork.
When I was with this teacher, I felt loved and cared
A84_BNS_Rel 628 1.0 6.0 3.62 1.38
about.
When I was with this teacher, I often felt a lot of
A85_BNS_Rel X 628 1.0 6.0 4.20 1.34
distance in our relationship.
When I was with this teacher, I felt a lot of
A86_BNS_Rel 628 1.0 6.0 2.55 1.34
closeness and intimacy.
A87_YCATS_W This teacher liked my family. 628 1.0 6.0 3.99 1.33
A88_YCATS_W This teacher liked me. 628 1.0 6.0 4.61 1.18
75
Table 2. Continued
A89_YCATS_W This teacher was my friend. 628 1.0 6.0 3.44 1.45
A90_YCATS_W This teacher said nice things about my work. 628 1.0 6.0 4.53 1.21
A91_YCATS_W This teacher helped me when I didn’t understand. 628 1.0 6.0 4.77 1.25
A92_YCATS_W This teacher made the class fun. 628 1.0 6.0 4.46 1.54
A93_YCATS_W This teacher answered my questions. 628 1.0 6.0 4.95 1.11
A94_YCATS_W This teacher told good stories. 628 1.0 6.0 4.55 1.45
A95_YCATS_W This teacher remembered special days for me. 628 1.0 6.0 2.90 1.41
A96_YCATS_W This teacher chose me to be a special helper. 628 1.0 6.0 2.57 1.30
A97_YCATS_W This teacher smiled a lot. 628 1.0 6.0 4.66 1.36
This teacher let me choose work that I wanted to
A98_YCAT_A 628 1.0 6.0 2.84 1.31
do.
A99_YCAT_A This teacher let me choose where I wanted to sit. 628 1.0 6.0 4.28 1.63
A100_YCAT_A This teacher let me do activities I wanted to do. 628 1.0 6.0 3.37 1.34
A101_YCAT_A This teacher let me do different activities in class. 628 1.0 6.0 3.24 1.43
A102_YCAT_A This teacher did activities with me. 628 1.0 6.0 3.61 1.38
A103_YCAT_A This teacher told me I was smart. 628 1.0 6.0 4.24 1.38
This teacher told me I was going to get in trouble a
A104_YCAT_C X 628 1.0 6.0 5.33 1.04
lot.
This teacher told me to do work that was too hard
A105_YCAT_C X 628 1.0 6.0 4.45 1.24
for me.
A106_YCAT_C This teacher got angry with me. X 628 1.0 6.0 5.01 1.21
This teacher told me that I was doing something
A107_YCAT_C X 628 1.0 6.0 4.79 1.18
wrong a lot.
A108_YCAT_C This teacher told me that I didn’t try hard enough. X 628 1.0 6.0 4.86 1.30
This teacher told me to stop doing work I liked
A109_YCAT_C X 628 1.0 6.0 5.15 0.93
doing.
A110_YCAT_C This teacher told me to do work I didn’t want to do. X 628 1.0 6.0 3.62 1.38
A111_YCAT_C This teacher told me I did not listen. X 628 1.0 6.0 5.04 1.12
A112_YCAT_C This teacher was mean. X 628 1.0 6.0 5.09 1.23
A113_YCAT_C This teacher had too many rules for our class. X 628 1.0 6.0 4.79 1.10
A114_STRS I trusted this teacher. 628 1.0 6.0 4.62 1.32
A115_STRS This teacher was in tune with what I was feeling. 628 1.0 6.0 3.55 1.31
A116_STRS I was wary of this teacher. X 628 1.0 6.0 4.67 1.32
A117_STRS I allowed this teacher to connect/relate with me. 628 1.0 6.0 3.91 1.32
I tried to get help, recognition, and support from
A118_STRS 628 1.0 6.0 3.93 1.27
this teacher.
This teacher was able to comfort me or cheer me
A119_STRS 628 1.0 6.0 3.76 1.38
up.
A120_STRS I avoided contact with this teacher. X 628 1.0 6.0 4.62 1.46
I tried to get support and encouragement from this
A121_STRS 628 1.0 6.0 3.79 1.31
teacher.
I saw this teacher as a source of punishment and
A122_STRS X 628 1.0 6.0 5.05 1.20
criticism.
A123_STRS I participated deeply in this teacher’s class. 628 1.0 6.0 4.20 1.32
A124_STRS I felt safe and secure with this teacher. 628 1.0 6.0 4.47 1.29
This teacher shared an affectionate, warm
A125_STRS 628 1.0 6.0 3.27 1.47
relationship with me.
A126_STRS If upset, I would seek comfort from this teacher. 628 1.0 6.0 2.98 1.43
A127_STRS I felt upset when separated from this teacher. 628 1.0 6.0 2.01 1.10
My relationship with this teacher became more
A128_STRS 628 1.0 6.0 4.29 1.39
positive throughout the school year.
76
Table 2. Continued
My relationship with this teacher became more
A129_STRS X 628 1.0 6.0 4.91 1.37
negative throughout the school year.
In my relationship with this teacher, I felt
A130_NRS_A 628 1.0 6.0 4.29 1.31
supported.
In my relationship with this teacher, I felt
A131_NRS_A 628 1.0 6.0 4.14 1.28
understood.
In my relationship with this teacher, I felt listened
A132_NRS_A 628 1.0 6.0 4.34 1.27
to.
A133_NRS_A In my relationship with this teacher, I felt valued. 628 1.0 6.0 4.22 1.30
A134_NRS_A In my relationship with this teacher, I felt safe. 628 1.0 6.0 4.58 1.20
In my relationship with this teacher, I felt close to
A135_NRS_I 628 1.0 6.0 3.46 1.42
them.
In my relationship with this teacher, I felt attached
A136_NRS_I 628 1.0 6.0 2.96 1.42
to them.
In my relationship with this teacher, I felt bonded
A137_NRS_I 628 1.0 6.0 3.39 1.44
to them.
In my relationship with this teacher, I felt close-
A138_NRS_I 628 1.0 6.0 3.30 1.44
knit.
In my relationship with this teacher, I felt like a
A139_NRS_I 628 1.0 6.0 3.41 1.48
friend.
If I had a problem with my school work I would
A140_Sch_Ut 628 1.0 6.0 4.02 1.44
share it with this teacher.
If I were having trouble understanding a subject at
A141_Sch_Ut 628 1.0 6.0 3.65 1.46
school, I would talk it over with this teacher.
I could usually rely on this teacher when I had
A142_Sch_Ut 628 1.0 6.0 3.67 1.48
problems at my school.
A143_Emul I tried to model myself after this teacher. 628 1.0 6.0 3.18 1.44
I would feel good if someone said I was a lot like
A144_Emul 628 1.0 6.0 3.77 1.56
this teacher.
A145_Emul I wanted to be like this teacher. 628 1.0 6.0 3.57 1.55
As I get older, I’ll probably be a lot like this
A146_Emul 628 1.0 6.0 3.28 1.49
teacher.
A147_ClsSup When I did well, this teacher made me feel good. 628 1.0 6.0 4.47 1.22
A148_ClsSup This teacher’s was like family to me. 628 1.0 6.0 2.67 1.45
A149_ClsSup This teacher was mean to me. X 628 1.0 6.0 5.22 1.16
A150_ClsSup This teacher just looked out for his/herself. X 628 1.0 6.0 4.90 1.25
When I had trouble with my schoolwork, this
A151_ClsSup 628 1.0 6.0 4.56 1.27
teacher tried to help.
A152_ClsSup This teacher worked with me to solve problems. 628 1.0 6.0 4.45 1.25
A153_ClsSup This teacher cared about me. 628 1.0 6.0 4.35 1.26
A154_ClsSup This teacher treated me with respect. 628 1.0 6.0 4.89 1.22
A155_ClsSup This teacher helped me, even if he/she was busy. 628 1.0 6.0 4.28 1.30
A156_ClsSup This teacher helped me learn. 628 1.0 6.0 4.81 1.24
This teacher let me help decide what the rules were
A157_MePart 628 1.0 6.0 2.52 1.21
going to be.
A158_MePart This teacher let me do things that I wanted to do. 628 1.0 6.0 3.45 1.34
This teacher let me get a rule changed if I thought
A159_MePart 628 1.0 6.0 2.78 1.28
it was unfair.
This teacher let me have a say in deciding what
A160_MePart 628 1.0 6.0 2.77 1.27
went on in class.
This teacher let me help plan what we would do in
A161_MePart 628 1.0 6.0 2.48 1.24
class.
A162_MePart This teacher let me do things my own way. 628 1.0 6.0 3.50 1.33
77
Table 2. Continued
In this class the teacher was the only one who
A163_MePart X 628 1.0 6.0 2.89 1.31
decided on the rules.
This teacher included me when planning what we
A164_MePart 628 1.0 6.0 2.73 1.29
did in class.
This teacher asked me to help decide what the class
A165_MePart 628 1.0 6.0 2.56 1.21
should do.
A166_MePart This teacher let me choose what I would work on. 628 1.0 6.0 2.96 1.31
A167_RAPS_E When I was with this teacher, I felt good. 628 1.0 6.0 4.06 1.35
A168_RAPS_E When I was with this teacher, I felt mad. X 628 1.0 6.0 5.00 1.25
A169_RAPS_E When I was with this teacher, I felt unhappy. X 628 1.0 6.0 4.85 1.27
A170_RAPS_E When I was with this teacher, I felt happy. 628 1.0 6.0 4.07 1.38
A171_Add_In I believe learning the content of this teacher’s class
628 1.0 6.0 4.74 1.22
is important.
A172_Add_In I felt that completing the assignments in this class
628 1.0 6.0 4.77 1.20
was beneficial to me.
A173_Add_In I believe the content of this teacher’s class is
628 1.0 6.0 4.73 1.21
valuable to understand.
A174_Add_In Because of this teacher, I believe more in the
628 1.0 6.0 4.13 1.50
purpose of learning the content of the class.
A175_Add_En I tried hard to do well in this teacher’s class. 628 1.0 6.0 4.93 1.13
A176_Add_En I paid attention when in this teacher’s class. 628 1.0 6.0 4.85 1.08
A177_Add_En I felt interested when in this teacher’s class. 628 1.0 6.0 4.29 1.47
A178_Add_En I enjoyed learning in this teacher’s class. 628 1.0 6.0 4.49 1.49
A179_Add_En I felt engaged by the content when in this teacher’s
628 1.0 6.0 4.33 1.41
class.
78
Confirmatory Factor Analyses of Original Fourteen Scales
Using SPSS AMOS Version 23, confirmatory factor analyses were performed to
examine each of the original fourteen relationship measures independently. Fit of the
observed data to the models implied by the measures were assessed with the χ2 statistic,
which reflects the absolute difference between actual and model-generated data. Because
the χ2 statistic is sensitive to sample size and this study has a fairly large sample (N =
628), the root mean square error of approximation (RMSEA) and its 90% confidence
interval are reported to provide another goodness of fit index that takes both sample size
and model complexity into account (Browne & Cudeck, 1993). In addition, the
comparative fit index (CFI), which has been shown to be one of the best incremental fit
indices (Bentler, 1990), is also reported to assess fit based on the assumption that model
may only provide a reasonable approximation of the population covariance matrix, which
is most likely the case here as it is in most social science research. RMSEA values less
than .08 and .05 suggest the model fit is adequate and close, respectively (Marsh, Hau, &
Wen, 2004). RMSEA values of .08 to .10 suggest a mediocre fit, whereas values .10 and
greater suggest a poor fit (Browne & Cudeck, 1993). CFI values in the range of .90 or
greater are generally considered to indicate reasonable model fit (Hu & Bentler, 1999)
and the criterion for a good fit is a CFI value of .95 or greater (Russell, 2002).
Standardized factor loadings (which represent the correlation between latent variable and
measured items), squared multiple correlations (which represent the variance in each item
accounted for by the latent variable), and Cronbach’s alpha coefficients (representing
internal consistency) are reported for each of the original fourteen scales under their sub-
79
headings below. A summary of the fit indices for each of the fourteen original scales is
Table 3. CFA model fit statistics for the fourteen original scales
RMSEA
Scale Title df χ2 p RMSEA CFI
CI (90%)
Classroom Life Measure 19 102.4 < .001 .084 .068 - .100 .978
Student Classroom Environment 14 39.9 < .001 .054 .035 - .075 .980
Measure
Developmental Studies Center 169 779.5 < .001 .076 .071 - .081 .929
Student Questionnaire
Inventory of Parent & Peer 272 1403.9 < .001 .081 .077 - .086 .912
Attachments
School Utilization & Teacher 13 30.6 .004 .046 .025 - .068 .995
Emulation
Research Assessment Package 2 135.3 < .001 .326 .281 - .374 .934
for Schools
Teacher as Social Context 71 548.2 < .001 .104 .096 - .112 .936
Student-Teacher Relationship 104 1170.9 < .001 .128 .121 - .135 .849
Scale
Young Children’s Appraisals 321 1712.8 < .001 .083 .079 - .087 .860
of Teacher Support
Psychological Sense of School 135 717.9 < .001 .083 .077 - .089 .942
Membership
Classroom Life Measure. As shown in Table 3, the goodness of fit statistics for
this measure are χ2(df = 19, N = 628) = 102.4, p < .001; CFI = .978; RMSEA = .084;
80
RMSEA 90% CI = .068 - .100. Standardized parameter estimates for the two-factor
model of the Classroom Life Measure items, consisting of the social support and the
academic support factors range from .62 to .90 and are Table 4. SMC estimates for
Classroom Life Measure
provided in Figure 3. The 4-item social support factor
Item Estimate
and the 4-item academic support factor coefficient A76_CLM_SS .807
alphas are .86 and .90 respectively. The estimated A77_CLM_SS .379
A78_CLM_SS .551
correlation between social support and academic A79_CLM_SS .795
support is .92. The squared multiple correlation A80_CLM_AS .775
A81_CLM_AS .614
(SMC) estimates range from .379 to .807 and are A82_CLM_AS .782
A83_CLM_AS .634
listed in Table 4.
fit statistics for this measure are χ2(df = 14, N = 628) = 39.9, p < .001; CFI = .980;
81
RMSEA = .054; RMSEA 90% CI = .035 - .075. Standardized parameter estimates for the
This 7-item factor had a coefficient alpha of .76. The Table 5. SMC estimates for
Student Classroom
squared multiple correlation (SMC) estimates ranged Environment Measure items
estimates for the two-factor model of the Developmental Studies Center Student
82
participation factors range from .58 to .87 and are displayed in Figure 5. The 10-item
classroom supportiveness factor and the 10-item meaningful participation factor have
83
Inventory of Parent & Peer Attachments. As Table 6. Continued
A157_MePart .517
shown in Table 3, the goodness of fit statistics for this A158_MePart .483
A159_MePart .449
measure are χ2(df = 272, N = 628) = 1403.9, p < .001; A160_MePart .667
A161_MePart .607
CFI = .912; RMSEA = .081; RMSEA 90% CI = .077
A162_MePart .452
- .086. Standardized parameter estimates for the A163_MePart .347
A164_MePart .514
three-factor model of the Inventory of Parent and Peer A165_MePart .582
A166_MePart .533
Attachments items, consisting of the trust,
84
communication, and alienation factors range from .46 Table 7. SMC estimates for
Inventory of Parent and Peer
to .87 and are displayed in Figure 6. The 10-item trust Attachments items
Item Estimate
factor, the 9-item communication factor, and the 6- A1_IPPA_T .750
A2_IPPA_T .723
item alienation factor have coefficient alphas of .95, A3_IPPA_T .726
A4_IPPA_T .747
.92, and .86 respectively. The estimated correlations A5_IPPA_T .220
A6_IPPA_T .684
between latent variables are .89 for trust and
A7_IPPA_T .656
communication, .93 for trust and alienation, and .77 for A8_IPPA_T .748
A9_IPPA_T .649
communication and alienation. The squared multiple A10_IPPA_T .742
A11_IPPA_C .609
correlation (SMC) estimates range from .214 to .748 A12_IPPA_C .525
A13_IPPA_C .615
and are listed in Table 7. A14_IPPA_C .214
A15_IPPA_C .605
School Utilization and Teacher Emulation. As
A16_IPPA_C .463
shown in Table 3, the goodness of fit statistics for this A17_IPPA_C .611
A18_IPPA_C .728
measure are χ2(df = 13, N = 628) = 30.6, p < .005; CFI A19_IPPA_C .682
A20_IPPA_A .505
= .995; RMSEA = .046; RMSEA 90% CI = .025 - A21_IPPA_A .644
A22_IPPA_A .281
.068. Standardized parameter estimates for the two- A23_IPPA_A .740
A24_IPPA_A .511
factor model of School Utilization and Teacher
A25_IPPA_A .488
Emulation items range from .75 to .92 and are displayed in Figure 7. The 3-item school
85
Figure 7. Standardized CFA parameter estimates of School Utilization and Teacher
Emulation items
fit statistics for this measure are χ2(df = 2, N = 628) = 135.3, p < .001; CFI = .934;
RMSEA = .326; RMSEA 90% CI = .281 - .374. Standardized parameter estimates for
the emotional security items from the Research Assessment Package for Schools range
from .83 to .89 and are displayed in Figure 8. This 4-item emotional security factor has a
86
coefficient alpha of .92. The squared multiple Table 9. SMC estimates for
Research Assessment Package
correlation (SMC) estimates range from .687 to .788 for Schools items
Item Estimate
and are listed in Table 9. A167_RAPS_E .763
A168_RAPS_E .687
Teacher as Social Context. As shown in
A169_RAPS_E .788
Table 3, the goodness of fit statistics for this measure A170_RAPS_E .770
are χ2(df = 71, N = 628) = 548.2, p < .001; CFI = .936; RMSEA = .104; RMSEA 90% CI
= .096 - .112. Standardized parameter estimates for the four-factor model of the Teacher
and dependability factors range from .55 to .89 and are displayed in Figure 9. The 3-item
affection factor, the 3-item attunement factor, the 2-item dedication of resources factor,
and the 6 item dependability factors have coefficient Table 10. SMC estimates for
Teacher as Social Context items
alphas of .88, .86, .77, and .90 respectively. The Item Estimate
A26_Inv_Aff .690
estimated correlations between latent variables are
A27_Inv_Aff .797
.90 for affection and attunement, .96 for affection A28_Inv_Aff .606
A29_Inv_Att .681
and dedication of resources, .96 for affection and A30_Inv_Att .776
A31_Inv_Att .630
dependability, .91 for attunement and dedication of A32_Inv_DR .612
A33_Inv_DR .641
resources, .90 for attunement and dependability, and A34_Inv_Dep .769
A35_Inv_Dep .776
.96 for dedication of resources and dependability.
A36_Inv_Dep .763
The squared multiple correlation (SMC) estimates A37_Inv_Dep .715
A38_Inv_Dep .401
range from .301 to .797 and are listed in Table 10. A39_Inv_Dep .301
87
Figure 9. Standardized CFA parameter estimates of Teacher as Social Context items
Sense of Relatedness. As shown in Table 3, the goodness of fit statistics for this
measure are χ2(df = 2, N = 628) = 34.9, p < .001; CFI = .981; RMSEA = .162; RMSEA
88
multiple correlation (SMC) estimates range from .560 to .788 and are listed in Table 11.
statistics for this measure are χ2(df = 14, N = 628) = 85.3, p < .001; CFI = .984; RMSEA
= .090; RMSEA 90% CI = .072 - .109. Standardized parameter estimates for the
Emotional Reliance Questionnaire items range from .84 to .91 and are displayed in
Figure 11. This 7-item emotional reliance factor has a coefficient alpha of .96. The
89
squared multiple correlation (SMC) estimates range .706 to Table 12. SMC estimates
for Emotional Reliance
.825 and are in Table 12. Questionnaire items
Item Estimate
Basic Need Satisfaction in Relationships. As shown A40_ERQ .766
A41_ERQ .825
in Table 3, the goodness of fit statistics for this measure are A42_ERQ .785
A43_ERQ .817
χ2(df = 0, N = 628) = 0.0, p = NA; CFI = 1.00; RMSEA = A44_ERQ .737
A45_ERQ .766
.567; RMSEA 90% CI = .530 - .606. Standardized A46_ERQ .706
parameter estimates for the relatedness need support items of the Basic Need Satisfaction
in Relationships scale range from .66 to .92 and are displayed in Figure 12. This 3-item
χ2(df = 34, N = 628) = 62.6, p < .005; CFI = .995; RMSEA = .037; RMSEA 90% CI =
.022 - .051. Standardized parameter estimates for the two-factor Need for Relatedness
Scale items, consisting of the acceptance and intimacy factors range from .73 to .92 and
are displayed in Figure 13. The 5-item acceptance factor and the 5-item intimacy factor
90
Figure 13. Standardized CFA parameter estimates of Need for Relatedness Scale items
have coefficient alphas of .94 and .93 respectively. Table 14. SMC estimates for
Need for Relatedness Scale items
The estimated correlation between acceptance and Item Estimate
A130_NRS_A .829
intimacy is .88. The squared multiple correlation A131_NRS_A .820
A132_NRS_A .809
(SMC) estimates range from .535 to .842 and are A133_NRS_A .812
A134_NRS_A .535
listed in Table 14.
A135_NRS_I .842
Student-Teacher Relationship Scale. As A136_NRS_I .587
A137_NRS_I .764
shown in Table 3, the goodness of fit statistics for A138_NRS_I .819
A139_NRS_I .682
this measure are χ2(df = 104, N = 628) = 1170.9, p <
.001; CFI = .849; RMSEA = .128; RMSEA 90% CI = .121 - .135. Standardized
parameter estimates for the Student-Teacher Relationship Scale items range from .37 to
.85 and are displayed in Figure 14. The 16-item relationship quality factor has a
91
Figure 14. Standardized CFA parameter estimates of Student-Teacher Relationship Scale
items
92
Table 15. Continued
Young Children’s Appraisals of
A122_STRS .485
Teacher Support. As shown in Table 3, the A123_STRS .341
A124_STRS .662
goodness of fit statistics for this measure are A125_STRS .510
A126_STRS .561
χ2(df = 321, N = 628) = 1712.8, p < .001; CFI A127_STRS .138
A128_STRS .676
= .860; RMSEA = .083; RMSEA 90% CI = A129_STRS .585
.079 - .087. Standardized parameter estimates
Table 16. SMC estimates for Young
for the three-factor model of the Young Children's Appraisals of Teacher
Support items
Children’s Appraisals of Teacher Support
Item Estimate
items, consisting of the warmth, the autonomy, A87_YCATS_W .385
A88_YCATS_W .666
and the conflict factors range from .37 to .84 A89_YCATS_W .584
A90_YCATS_W .689
and are displayed in Figure 15. The 11-item A91_YCATS_W .671
A92_YCATS_W .707
warmth factor, the 6-item autonomy factor, and A93_YCATS_W .587
A94_YCATS_W .559
the 10-item conflict factor have coefficient
A95_YCATS_W .335
alphas of .92, .79 and .88 respectively. The A96_YCATS_W .210
A97_YCATS_W .534
estimated correlations between latent variables A98_YCAT_A .443
A99_YCAT_A .202
are .87 for warmth and autonomy, .80 for A100_YCAT_A .552
A101_YCAT_A .334
warmth and conflict, and .60 for autonomy and A102_YCAT_A .448
A103_YCAT_A .474
conflict. The squared multiple correlation
A104_YCAT_C .530
(SMC) estimates range from .135 to .689 and A105_YCAT_C .146
A106_YCAT_C .641
are listed in Table 16. A107_YCAT_C .561
A108_YCAT_C .558
A109_YCAT_C .429
A110_YCAT_C .135
A111_YCAT_C .493
A112_YCAT_C .658
A113_YCAT_C .430
93
Figure 15. Standardized CFA parameter estimates of Young Children’s Appraisals of
Teacher Support items
94
Psychological Sense of School Membership. As shown in Table 3, the goodness of
fit statistics for this measure are χ2(df = 135, N = 628) = 717.9, p < .001; CFI = .942;
RMSEA = .083; RMSEA 90% CI = .077 - .089. Standardized parameter estimates for
the Psychological Sense of School Membership items range from .64 to .87 and are
displayed in Figure 16. This 18-item belongingness factor has a coefficient alpha of .97.
95
The squared multiple correlation (SMC) estimates range Table 17. SMC estimates for
Psychological Sense of
from .408 to .755 and are listed in Table 17. School Membership items
Item Estimate
Regression of Scale Scores on Internalization, A58_PSSM .747
A59_PSSM .637
Engagement, and Academic Achievement A60_PSSM .713
A61_PSSM .591
To address the second hypothesis, scale scores
A62_PSSM .508
were calculated for each of the original scales. Scale A63_PSSM .629
A64_PSSM .694
scores were also calculated for the internalization and A65_PSSM .625
A66_PSSM .527
engagement measures; whereas, academic achievement A67_PSSM .753
A68_PSSM .408
was only measured with one item so scale score A69_PSSM .656
A70_PSSM .752
calculation was not necessary. Next, linear regression
A71_PSSM .470
was used to examine the predictive validity of each of A72_PSSM .702
A73_PSSM .697
the original scales in relation to each of the outcomes, A74_PSSM .755
A75_PSSM .548
internalization, engagement, and academic achievement.
Calculating scale scores. For each of the 628 participants, scale scores for the
original fourteen instruments were calculated using SPSS version 22. If the original
instrument was divided into more than one factor by its original author, scores were
calculated for each of those factors, also referred to as subscale scores. For instance, the
Psychological Sense of School Membership scale consists of only one factor called
belongingness (Figure 16); therefore, only one score was calculated for that scale for each
respondent. Whereas, the Teacher as Social Context scale produces four factors, which
are titled affection, attunement, dedication of resources, and dependability (Figure 9);
therefore, four subscale scores, one for each of the factors, were calculated for each
respondent. The calculation of all scale and subscale scores for each participant consisted
96
of adding the scores for items included in the scale/subscale and dividing the sum by the
number of items in the scale/subscale. In other words, scale scores are an average of the
scores for the items included in the factor. Even though several of the factor loadings and
squared multiple correlations, displayed in Figures 3-16 and Tables 4-17 respectively,
suggest that the inclusion of some items is questionable, all of the original items were
Although items from several of the original fourteen scales are hypothesized to fit well
into a model that has them directly loading on the latent variables of internalization and
engagement, those items were not included in the calculation of subscale scores for
internalization and engagement for the purposes of examining the predictive validity of
the original fourteen instruments. Instead only the 4 additional internalization items and
the 5 additional engagement items were considered for the calculation of scores for
internalization and engagement. Only one C-STRI item measured academic achievement
by asking participants to select the letter grade they received in the class. Because higher
scores on the relationship quality items represent higher quality relationships, the letter
C = 6, C- = 5, D+ = 4, D = 3, D- = 2, F = 1.
Before calculating the subscale scores for internalization and engagement, the
internal consistency, factor loadings, and model fit were examined using Confirmatory
Factor Analysis. As displayed in Figure 17, the factor loadings of the four internalization
97
Figure 17. Standardized CFA parameter estimates of internalization items added to the C-
STRI for examining predictive validity of original scales
This 4-item internalization factor has acceptable internal consistency (alpha = .87) with a
χ2(df = 2, N = 628) = 22.86, p < .001; CFI = .984; RMSEA = .129; RMSEA 90% CI =
.085 - .179. Although the CFI suggests a good fit, the RMSEA indicates a poor fit
between observed data and the model. Therefore, the modification indices and factor
loading estimates were reviewed for ways to improve the model. The factor loadings of
the unrevised model range from .71 to .87 so all four items seem to belong as indicators
for the latent variable internalization. The modification indices provided by AMOS
indicate a modest but existing covariance between the error terms e2 and e4 (M.I. =
19.9). Adding the covariance path between e2 and e4 to the model left only one degree
of freedom, therefore, error variance parameter estimates were examined. The estimated
variance of e1 = .346 and the estimated variance of e3 = .357; therefore these parameters
were constrained to be equal in order to add a degree of freedom for model identification.
98
Figure 18. Standardized CFA parameter estimates for revised model of internalization
items added to the C-STRI for examining predictive validity of original scales
For this revised model of the internalization items, the fit statistics were χ2(df = 2, N =
628) = .582, p = .748; CFI = 1.0; RMSEA = .000; RMSEA 90% CI = .000 - .055, which
suggest good fit between observed data and the model. The five items added to the C-
STRI to measure engagement were examined next. The factor loading estimations are
Figure 19. Standardized CFA parameter estimates of engagement items added to the C-
STRI for examining predictive validity of original scales
This 5-item engagement factor showed acceptable internal consistency (alpha = .88);
however, the model fit statistics suggested this model required modifications in order to
obtain acceptable fit (χ2 = 144.34, df = 5, N = 628, p < .001; CFI = .930; RMSEA = .211;
99
RMSEA 90% CI = .182 - .241). The modification indices provided by AMOS indicated
strong covariance between the error terms e1 and e2 (M.I. = 115.9). Before adding that
covariance path to the model, the factor loading estimates were examined for any items
lower than .60 that might need to be deleted from the model. The estimated factor
loading of .52 for item number 175 on the latent variable engagement suggested it as
good candidate for deletion. Thus, item 175 was deleted and the new parameter estimates
as well as model fit were again examined. Figure 20 shows the estimated factor loadings
Figure 20. Standardized CFA parameter estimates of the 4-item model of engagement
items added to the C-STRI for examining predictive validity of original scales
This revised, 4-item model of the engagement’s fit statistics were χ2(df = 2, N = 628) =
7.30, p = .026; CFI = .997; RMSEA = .065; RMSEA 90% CI = .019 - .119, which
In accordance with the CFA analyses, the four internalization items were added
together and divided by 4 to create an internalization scale score for each participant.
Similarly, the four engagement items that remained in the reduced 4-item model were
added together and divided by 4 to create an engagement scale score for each participant.
100
Regression of original scales on internalization. Using participants’
internalization scale score as the dependent variable, fourteen separate linear regression
analyses were performed, one for each original instrument. The subscale score(s) were
entered as the independent variable(s). For instruments that had more than one subscale,
the Entry method in SPSS version 22 was used to include all subscale scores in the
analyses, which forces all independent variables into the equation in one step. The results
Table 18. Regression results for each original scale with internalization as the dependent
variable
Scale Subscale B SE B βeta R2
Social Support .119 .048 .120*
Classroom Life Measure .530***
Academic Support .677 .052 .626***
Student Classroom
Friendly/Fair .687 .045 .518*** .268***
Environment Measure
Classroom
.738 .036 .692***
Developmental Studies Center Supportiveness
.498***
Student Questionnaire Meaningful
.029 .038 .026
Participation
101
Table 18. Continued
Emotional Reliance
NA .514 .025 .640*** .409***
Questionnaire
Student-Teacher Relationship
NA .797 .031 .713*** .507***
Scale
regressions were performed with participant’s engagement scale score used as the
Table 19. Regression results for each original scale with engagement as the dependent
variable
Scale Subscale B SE B βeta R2
Social Support .180 .047 .165***
Classroom Life Measure .621***
Academic Support .767 .051 .647***
Student Classroom
Friendly/Fair .849 .047 .584*** .341***
Environment Measure
Classroom
.860 .034 .737***
Developmental Studies Center Supportiveness
.614***
Student Questionnaire Meaningful
.103 .037 .082**
Participation
102
Table 19. Continued
Affection .305 .058 .283***
Attunement .106 .045 .113*
Teacher as Social Context
Dedication of .575***
.131 .048 .128**
Resources
Dependability .313 .059 .294***
Emotional Reliance
NA .623 .025 .708*** .501***
Questionnaire
Student-Teacher Relationship
NA .968 .030 .790*** .624***
Scale
fourteen regressions were performed, using the participant’s self-reported grade they
achieved in the class as the dependent variable. The results are presented in Table 20.
Table 20. Regression results for each original scale with academic achievement as the
dependent variable
Scale Subscale B SE B βeta R2
Social Support .171 .111 .094
Classroom Life Measure .235***
Academic Support .795 .120 .404***
Student Classroom
Friendly/Fair 1.080 .086 .448*** .201***
Environment Measure
Classroom
.729 .081 .376***
Developmental Studies Center Supportiveness
.216***
Student Questionnaire Meaningful
.298 .087 .139**
Participation
103
Table 20. Continued
Trust .863 .148 .472***
Inventory of Parent & Peer
Communication -.217 .116 -.117 .233***
Attachments
Alienation .229 .127 .119
Emotional Reliance
NA .589 .053 .403*** .162***
Questionnaire
Student-Teacher Relationship
NA .943 .072 .464*** .215***
Scale
examine the measurement model of C-STRI items as organized into the factors displayed
in Figure 1. To simplify this portion of the analysis, due to the massive number of
predictor items, the model was initially divided into five separate parts for independent
analyses. The four predictor variables, academic support, punishment, involvement, and
104
autonomy support made up the first part. The second part consisted of the factors caring
and awareness/understanding, which are part of the higher order latent variable
perceptions of teacher. Next, the higher order latent variable perceptions of self with its
closeness and affective experience factors made up the third part. The fourth part was the
higher order latent variable interpersonal scripts along with its social support seeking and
emotional reliance factors. Finally, the fifth part of the model examined included the
five model parts were checked for model fit and modified until the model’s goodness of
fit statistics reached acceptable levels. After the five modified parts independently
produced acceptable model fit statistics, the full structural model was tested for goodness
of fit.
support. Figure 21 shows the initial confirmatory factor analysis model of all items
hypothesized to fit in these factors after the categorical item sorting process. This 4-
factor hypothesized model of all the predictor variable items produced the goodness of fit
statistics χ2(df = 2204, N = 628) = 7313.54, p < .001; CFI = .850; RMSEA = .061;
RMSEA 90% CI = .059 - .062. While the RMSEA value is in the acceptable range, the
CFI value suggests poor fit between observed data and the model. The first modification
to the hypothesized measurement model of predictor factors was made based on the latent
covariance of .97 suggests that the items hypothesized to be measuring the teacher’s
providing of academic support were possibly measuring the same construct as the items
105
Figure 21. Latent variable covariance and factor loading parameter estimates for the
hypothesized measurement model of predictor variables
106
hypothesized to be measuring the teacher’s involvement or appreciation, understanding
and respect for the student. Upon reviewing the content of the items it seems obvious in
hindsight that the behaviors and actions that relate to the concept of involvement for a
teacher would most often be academic support behaviors as that is the primary nature of a
teacher’s job. Therefore, the latent variable for academic support was eliminated and all
of its items were transferred to the involvement factor. Figure 22 displays the 3-factor
model with standardized parameter estimates and all of the items included. Goodness of
fit statistics for this model are χ2(df = 2207, N = 628) = 7485.97, p < .001; CFI = .846;
RMSEA = .062; RMSEA 90% CI = .060 - .063. Although model parsimony improved,
the fit statistics still suggest relatively poor model fit. In other words, the model required
further modifications. To improve the more parsimonious 3-factor model, due to the
extremely large number of items associated with each factor, item elimination was the
next technique used. First, all items with standardized factor loadings less than .60 were
Similarly, items A47_Friend and A52_Friend were removed from the punishment factor.
A163_MePart were removed from the autonomy-support factor. Table 21 lists the item
wording, standardized parameter estimate, and title of the latent variable hypothesized to
be measured by the item. Goodness of fit statistics for this 3-factor model with 12 items
removed are χ2(df = 1481, N = 628) = 5032.65, p < .001; CFI = .885; RMSEA = .062;
RMSEA 90% CI = .060 - .064. Although improved, the reduced item, 3-factor model
107
Figure 22. Standardized parameter estimates for 3-factor model of predictors with all
items included
108
Table 21. Items deleted from predictor factors in hypothesized model due to low factor
loadings
Standardized Associated
Item Wording Estimate Factor
A5_IPPA_T This teacher expected too much of me. (reverse score) .47 Involvement
I couldn’t count on this teacher when I needed him/her.
A39_Inv_Dep .55 Involvement
(reverse score)
A95_YCATS_W This teacher remembered special days for me. .56 Involvement
A96_YCATS_W This teacher chose me to be a special helper. .42 Involvement
This teacher told me to do work that was too hard for
A105_YCAT_C .31 Involvement
me. (reverse score)
This teacher criticized me for turning work in late or
A47_Friend .43 Punishment
failing to turn in assignments.
A52_Friend This teacher used sarcasm. .05 Punishment
Autonomy-
A99_YCAT_A This teacher let me choose where I wanted to sit. .45
Support
This teacher told me to stop doing work I liked doing. Autonomy-
A109_YCAT_C .31
(reverse score) Support
This teacher told me to do work I didn’t want to do. Autonomy-
A110_YCAT_C .25
(reverse score) Support
This teacher had too many rules for our class. (reverse Autonomy-
A113_YCAT_C .36
score) Support
In this class the teacher was the only one who decided Autonomy-
A163_MePart .58
on the rules. (reverse score) Support
The next strategy used to improve model fit for the items that remained in this
portion of the hypothesized model involved examination of cross loading estimates from
for items on latent variables other than the one they are hypothesized to correspond with
in the model were scanned for modification indices greater than 10.0. Using this
strategy, 15 more items were identified for elimination. Table 22 lists these 15 items,
their wording, the latent variable they were hypothesized to measure, and any
unacceptable (M.I. > 10.0) regression weight modification indices that they had for latent
109
Table 22. Items deleted from predictor factors in hypothesized model due to high cross
loading modification estimates
Autonomy-
Item Wording Involvement Punishment
Support
A32_Inv_DR This teacher spent time with me. Hyp 13.69 17.92
This teacher made sure I was included in
A66_PSSM Hyp* 11.88
lots of activities in class.
A67_PSSM I was treated with respect by this teacher. Hyp 13.02
A93_YCATS_W This teacher answered my questions. Hyp* 13.33
A102_YCAT_A This teacher did activities with me. Hyp* 37.36
My relationship with this teacher became
A129_STRS more negative throughout the school year. Hyp 25.05
(reverse score)
This teacher just looked out for his/herself.
A150_ClsSup Hyp 11.41
(reverse score)
A154_ClsSup This teacher treated me with respect. Hyp 12.33
This teacher threatened to give me more
A53_Friend work, tests, or to lower my grade to control Hyp 12.09
my behavior.
A120_STRS I avoided contact with this teacher. 28.51 Hyp 27.90
This teacher let me do activities I wanted
A100_YCAT_A 16.67 Hyp
to do.
This teacher let me help decide what the
A157_MePart 11.41 Hyp
rules were going to be.
This teacher let me do things that I wanted
A158_MePart 13.37 Hyp
to do.
This teacher let me help plan what we
A161_MePart 13.28 Hyp
would do in class.
A162_MePart This teacher let me do things my own way. 16.05 21.98 Hyp
*Originally hypothesized to be in the academic-support factor, but included in the involvement factor in the
revised 3-factor model
With these 15 items removed, the goodness of fit statistics are χ2(df = 776, N = 628) =
2408.62, p < .001; CFI = .924; RMSEA = .058; RMSEA 90% CI = .055 - .061. These fit
statistics suggest a reasonable, but not good fit; therefore, one more modification strategy
was used to further improve the fit. Specifically, error covariance modification indices
were examined for items within each latent variable. Because there is no magic number
for this kind of strategy, the modifications indices were examined to determine the size of
modification indices to act on by adding error covariance parameters to the model. The
five largest indices were greater than 60.0 and the sixth largest was less than 50.0 so the
model was modified once more by adding five error covariance parameters.
110
Table 23 displays the error terms for which covariance
Table 23. Five highest error
parameters were added as well as the associated covariance modification
indices for items in
modification indices that led to that modification. Figure hypothesized predictor
factor measurement model
Error Error M.I.
23 is a representation of the modified measurement model Label Label
e1 <--> e16 96.9
for predictor variables in the hypothesized structural e53 <--> e54 66.2
e32 <--> e64 66.1
model after the 12 items were removed for low factor e3 <--> e11 62.6
e28 <--> e31 62.0
loadings, the 15 items were removed for factor cross loadings, and the 5 strongest within-
factor item error covariance were added. The goodness of fit statistics χ2(df = 771, N =
628) = 2034.51, p < .001; CFI = .941; RMSEA = .051; RMSEA 90% CI = .048 - .054
now suggest good fit of observed data to the model, which has 23 items associated with
the involvement factor, 11 items associated with the punishment factor, and 7 items
111
Figure 23. Standardized parameter estimates for 3-factor model of predictors with items
removed and error covariance added
112
Perceptions of teacher. The hypothesized measurement model for the perceptions
labeled perceptions of teacher. The two first order latent variables are labeled
analysis model of all items hypothesized to load on these factors after the categorical item
sorting process.
Figure 24. Standardized parameter estimates for the hypothesized second-order factor
model of perceptions of teacher with all items included
113
The goodness of fit statistics χ2(df = 628, N = 628) = 3187.58, p < .001; CFI = .893;
RMSEA = .081; RMSEA 90% CI = .078 - .083 for this model suggest a borderline poor
fit of the observed data to the model. Due to the large number of items associated with
each factor in this portion of the hypothesized model, item elimination was a highly
desired technique for model fit improvement. However, due to the high factor loadings,
particularly for the caring factor, the criterion for item elimination was increased.
Specifically, items with standardized factor loadings of less than .65 were eliminated.
factor and the items A77_CLM_SS, A87_YCATS_W, and A148_ClsSup being removed
from the caring factor. Table 24 lists the item wording, standardized parameter estimate,
and title of the latent variable hypothesized to be measured by the item. The goodness of
fit statistics for this model reduced by 4 items χ2(df = 494, N = 628) = 2407.95, p < .001;
CFI = .914; RMSEA = .079; RMSEA 90% CI = .075 - .082 for this model suggest an
Table 24. Items deleted from perceptions of teacher measurement model due to low
factor loadings
Standardized
Item Wording Associated Factor
Estimate
A22_IPPA_A I get upset a lot more than this teacher .49 Awareness/Understanding
knew about. (reverse score)
A77_CLM_SS This teacher thought it was important to be .63 Caring
my friend.
A87_YCATS_W This teacher liked my family. .64 Caring
A148_ClsSup This teacher’s was like family to me. .64 Caring
Examination of cross loading estimates from the modification indices was again
utilized to determine the item elimination for improved model fit. This time, only one
friend,” had a cross loading modification indicator greater than 10.0 (M.I. = 13.95).
114
Eliminating this item negligibly improved the model fit to χ2(df = 463, N = 628) =
2190.02, p < .001; CFI = .920; RMSEA = .077; RMSEA 90% CI = .074 - .080. To reach
good model fit, error covariance modification indices were examined next. The five
largest indices were greater than 50.0; thus, the model was modified by adding five error
covariance parameters. Table 25 displays the error Table 25. Five highest error
covariance modification indices
terms for which covariance parameters were added as for items in hypothesized
perceptions of teacher portion of
well as the associated modification indices that led to the measurement model
Error Error
that modification. Figure 25 is a representation of the Label Label M.I.
e6 <--> e7 149.2
modified measurement model for the perceptions of e31 <--> e35 82.9
e27 <--> e29 75.1
e14 <--> e23 58.4
teacher factors in the hypothesized structural model
e13 <--> e33 51.9
after the 4 items were removed for low factor loadings, 1 item was removed for its factor
cross loading, and the 5 strongest within-factor item error covariance were added. The
goodness of fit statistics χ2(df = 458, N = 628) = 1750.35, p < .001; CFI = .940; RMSEA
= .067; RMSEA 90% CI = .064 - .070 now suggest a fairly good fit between the observed
data and the model, which has 8 items associated with the awareness/understanding factor
115
Figure 25. Standardized parameter estimates for the second-order perceptions of teacher
factor model with items removed and error covariance added
Perceptions of self. The next part of the model to examine, and improve if
116
hypothesized model with all items included is displayed in Figure 26 and has goodness of
fit statistics χ2(df = 349, N = 628) = 2835.76, p < .001; CFI = .869; RMSEA = .107;
RMSEA 90% CI = .103 - .110, which suggests the data do not fit the model. In other
words, the model needs modification before it can be used in testing the hypothesized
structural model.
Figure 26. Standardized parameter estimates for the hypothesized second-order factor
model of Perceptions of Self with all items included
117
Following the same strategies of model modification until a good fit is achieved, the first
the first-order factors to the second-order factors. That is, the .98 loading of Closeness
and the .96 loading of Affective Experience suggest that the items in this measurement
model of all the items in this portion of the hypothesized model resulted in a decrease of
model fit; specifically, χ2(df = 350, N = 628) = 3059.99, p < .001; CFI = .857; RMSEA =
.111; RMSEA 90% CI = .108 - .115. Thus, the single-factor model modification was
rejected. Next, item elimination began with items having standardized factor loadings
less than .60 removed. This step only removed item A127_STRS, worded “I felt upset
when separated from this teacher,” from the Affective Experience factor. This resulted in
a very slight improvement to the fit statistics χ2(df = 323, N = 628) = 2652.63, p < .001;
CFI = .876; RMSEA = .107; RMSEA 90% CI = .103 - .111, but the data still do not fit
the model. The next step in modification involved examining cross factor loading
modification indicators for values greater than 10.0, of which there were none. Finally,
error covariance modification indices were examined. The four largest indices, listed in
Table 26, were extremely high. The model was Table 26. Four highest error
covariance modification indices
modified by adding four error covariance parameters, for items in hypothesized
Perceptions of Self portion of
which improved the model fit χ2(df = 319, N = 628) = the measurement model
Error Error
2149.57, p < .001; CFI = .903; RMSEA = .096; Label Label M.I.
e17 <--> e20 173.6
RMSEA 90% CI = .092 - .100. Yet, because the e8 <--> e18 110.0
e17 <--> e19 108.4
e19 <--> e20 105.9
covariance parameter estimates were really high and
because three of the error terms included multiple covariance parameters, elimination of
118
the five items associated with these error terms was performed as the next step in
in further improvement of model fit χ2(df = 208, N = 628) = 1416.46, p < .001; CFI =
.921; RMSEA = .096; RMSEA 90% CI = .092 - .101. Although the model fit indices are
modification indices was performed. This time, using greater than 50 as the cutoff, just
as was done with the Perceptions of Teacher portion of the model, seven error covariance
parameters were added to the model. Table 27 Table 27. Seven remaining
allowable error covariance
shows the associated error covariance modification modification indices greater than
50 for items in the hypothesized
indices. Additionally, the error term e10 had four Perceptions of Self portion of the
measurement model
unallowable (error covariance with error terms not Error Error
Label Label M.I.
associated with items in the same factor) error e23 <--> e26 82.2
e22 <--> e26 80.8
e25 <--> e28 76.3
covariance modification indices greater than 60, so e1 <--> e9 73.8
e22 <--> e23 62.3
its associated item, A116_STRS was eliminated at e11 <--> e16 55.9
e25 <--> e26 55.1
this step as well. Figure 27 is a representation of the
modified measurement model for the Perceptions of Self portion of the hypothesized
structural model after the items were removed for low factor loadings and unacceptable
error covariance modification indices and seven within-factor item error covariance
parameters were added. The resulting goodness of fit statistics for this model χ2(df = 181,
N = 628) = 792.23, p < .001; CFI = .958; RMSEA = .073; RMSEA 90% CI = .068 - .079
now suggest a fairly good fit between the observed data and the model, which has 15
119
items associated with the closeness factor and 6 items associated with the affective
experience factor.
Figure 27. Standardized parameter estimates for the second-order Perceptions of Self
factor model with items removed and error covariance added
120
Interpersonal scripts. The interpersonal scripts portion of the hypothesized
interactions between themselves and their teacher, with 11 items associated with the
social support seeking factor and 10 items associated with emotional reliance is displayed
in Figure 28 and has goodness of fit statistics χ2(df = 188, N = 628) = 787.99, p < .001;
CFI = .952; RMSEA = .071; RMSEA 90% CI = .066 - .077, which suggests a fairly good
fit between the observed data and the model. In other words, it can be used as is when
Figure 28. Standardized parameter estimates for the hypothesized second-order factor
model of Interpersonal Scripts with all items included
121
However, the dual .99 standardized regression weights between the first-order and
second-order factors suggests that a single factor model might fit the data even better and
would be a more parsimonious model. The single-factor model of all the items in this
portion of the hypothesized model resulted in a decrease of model fit so it was rejected.
engagement, and academic achievement as outcome variables with all items included is
displayed in Figure 29. The goodness of fit statistics for this portion of the model are
Figure 29. Standardized parameter estimates for the hypothesized outcome variables with
all items included
122
χ2(df = 297, N = 628) = 2767.11, p < .001; CFI = .829; RMSEA = .115; RMSEA 90% CI
= .111 - .119, which suggests the data do not fit the model. That is, the model of items
measuring the outcome variables needs modification before it can be used in testing the
full structural model. Keeping with the model modification strategies utilized for the
other measurement models in in this study, item elimination began with items having
standardized factor loadings less than .60 removed. This step removed item
A118_STRS, A121_STRS, and A175_Add_En. Table 28 shows the wording for these
deleted items. This resulted in an improvement to the fit statistics χ2(df = 228, N = 628)
Table 28. Items deleted from the outcome variables’ measurement model due to low
factor loadings
Standardized
Item Wording Estimate Associated Factor
A118_STRS I tried to get help, recognition, and support .42 Engagement
from this teacher.
A121_STRS I tried to get support and encouragement .51 Engagement
from this teacher.
A175_Add_En I tried hard to do well in this teacher’s class. .48 Engagement
= 2150.81, p < .001; CFI = .857; RMSEA = .116; RMSEA 90% CI = .112 - .120, but the
data still do not fit the model. The next step in modification involved examining cross
factor loading modification indicators for values greater than 10.0, of which there were
the six displayed in Table 29. Whereas, this modification did result in a slight
improvement to the fit statistics χ2(df = 117, N = 628) = 1155.02, p < .001; CFI = .89;
RMSEA = .119; RMSEA 90% CI = .113 - .125, the data still do not fit the model.
Additionally, examination of the wording of items that have been removed from the
model and the items that are left in the model reveals that the fit is being improved by
moving away from internalization toward something more like emulation in the first
123
Table 29. Items deleted from outcome factors in hypothesized model due to high cross
loading modification estimates
Item Wording Internalization Engagement Achievement
A172_Add_In I felt that completing the assignments Hyp 16.86
in this class was beneficial to me.
A173_Add_In I believe the content of this teacher’s Hyp 12.69
class is valuable to understand.
A62_PSSM Sometimes I felt as if I didn’t belong Hyp 14.52
in this teacher’s class.
A92_YCATS_W This teacher made the class fun. Hyp 14.16
A123_STRS I participated deeply in this teacher’s Hyp 17.87
class.
A176_Add_En I paid attention when in this teacher’s Hyp 10.27
class.
factor. Because a set of items had been added that directly represented the factors of
variables’ measurement model to improve model fit was attempted. Starting again with
the items displayed in Figure 29, all items that came from the original scales were deleted
from the internalization factor, because they represented emulation or desired emulation
Next the same was done for the engagement factor because the items from the original
scales represented liking the teacher, belonging in class, and help seeking efforts as
opposed to direct statements about behavioral or emotional engagement in the class. This
improved the fit statistics to χ2(df = 33, N = 628) = 411.16, p < .001; CFI = .912; RMSEA
= .135; RMSEA 90% CI = .124 - .147, yet the model is still not acceptable. In an effort
to further improve the model fit, the standardized parameter estimates for the factor
loadings were examined. The lowest two were .54 for A175_Add_En and .68 for
A176_Add_En. Comparison between the content of these two Engagement items with
the content of the other three remaining Engagement factor items revealed that the two
with low loadings are related to behavioral aspects of Engagement and the other three are
124
related to emotional aspects of engagement. Thus, the next step in improving the fit of
this measurement model involved splitting those five Engagement items into two separate
improved the model’s goodness of fit χ2(df = 30, N = 628) = 270.05, p < .001; CFI =
.944; RMSEA = .113; RMSEA 90% CI = .101 - .126; however, model fit is not
acceptable. Additional modification was still required. To this end, error covariance
modification indices were examined. Only one pair of error terms had an error
covariance modification index greater than 50. The error term for A171_Add_In, e6, and
the error term for A173_Add_In, e8 had the modification index of 117.16 so their
covariance parameter was added to the model. The model after this modification, which
has 4 items associated with the internalization factor, 2 items for the behavioral
engagement factor, 3 items for the emotional engagement factor, and 1 item for the
academic achievement factor, produced acceptable goodness of fit statistics χ2(df = 29, N
= 628) = 119.64, p < .001; CFI = .979; RMSEA = .071; RMSEA 90% CI = .058 - .084
125
Figure 30. Standardized parameter estimates for the hypothesized outcome variables after
item elimination and splitting of behavioral and emotional engagement
To address the fourth hypothesis, the modified measurement model parts were
combined into a single model with the hypothesized latent variable associations, as
displayed in Figure 31. Unfortunately, the resultant model was found to have problems
of underidentification. In other words, the model contained fewer known parameters than
have been to search for a set of possibly justifiable restrictions to impose in the model in
order to reduce the number of unknown parameters. However, options within this
method seemed superfluous when used purely to reach identification in a complex model.
126
Figure 31. Full structural equation model with the hypothesized latent variable
associations using the modified measurement model parts
accomplish this, each of these variables was treated as a subscale by calculating simple
means of their respective items from the modified measurement model parts. This
greatly reduced the number of unknown parameters in the model resulting in the model
displayed in Figure 32. This model is overidentified. It has a greater number of known
to have an adequate test of model fit (Bolen, 1989). The model in Figure 32 produced
its goodness of fit statistics χ2 (df = 1522, N = 628) = 5977.80, p < .001; CFI = .878;
127
Figure 32. Full structural equation model with unknown parameters reduced by
calculating subscale scores for several factors from the modified measurement model
specifications
RMSEA = .068; RMSEA 90% CI = .067 - .070. To improve this model, the error
covariance and regression weight modification indices were examined. The regression
weight modification indices were first sorted alphabetically to remove all pathways
except those involving only latent variables. Next, they were sorted from highest to
lowest to determine what pathways might have sufficient empirical utility in order to be
considered for theoretical justification. The highest modification indicator was for the
pathway arguably makes more sense than the pathway from interpersonal scripts to
academic achievement in that relationship quality might have direct effects on the
student’s internalization and engagement, which then leads to more or less academic
achievement accordingly. In other words, modification of the model using this logic
128
would require the removal of the pathway from interpersonal scripts to academic
achievement and the addition of three paths from behavioral engagement to academic
achievement. Table 30 lists the modification indices for these three regression pathways.
The resultant model demonstrated improved goodness of fit statistics χ2(df = 1520, N =
628) = 5493.79, p < .001; CFI = .891; RMSEA = .065; RMSEA 90% CI = .063 - .066
and is displayed in Figure 33. None of the remaining regression weight modification
indices of substantive size presented any path modifications that could be justified with
Figure 33. Full model using subscale scores and treating academic achievement as a
secondary outcome
129
theory. Therefore, in order to improve the model further, error covariance modification
To accomplish this, the error covariance modification indices were sorted to retain
only the ones related to covariance between latent variable error terms of the same level
in the model and covariance between error terms associated with subscale scores
measuring the same latent variable. Of these error covariance parameters, the top four
sense that these latent variables are interrelated, regression lines connecting them were
not a desired part of this model. Given that, it made theoretical sense to add covariance
The other modification indicator in table 30, the second highest, was for the
covariance parameter between the error terms associated with the subscales emotional
reliance and social support seeking. This parameter was also acceptable and added to the
model. The modified model is shown in Figure 34. This version produced improved
goodness of fit statistics χ2(df = 1516, N = 628) = 4755.03, p < .001; CFI = .911; RMSEA
= .058; RMSEA 90% CI = .057 - .060. With an adequate RMSEA and a reasonable CFI,
the model in Figure 34 was acceptable. The parameter estimates for factor loadings and
130
Figure 34. Full model using subscale scores and treating academic achievement as a
secondary outcome and modified by the addition of error covariance parameters
Table 32. Parameter estimates for factor loadings and path coefficients in the final
version of the full model
Latent Variables Indicators Parameter Estimates
Factor Loadings
Involvement A2_IPPA_T 0.843
A7_IPPA_T 0.820
A24_IPPA_A 0.753
A33_Inv_DR 0.787
A34_Inv_Dep 0.859
A35_Inv_Dep 0.854
A36_Inv_Dep 0.841
A37_Inv_Dep 0.847
A38_Inv_Dep 0.632
A56_Rel 0.837
A59_PSSM 0.809
A60_PSSM 0.847
A71_PSSM 0.685
A90_YCATS_W 0.818
A91_YCATS_W 0.828
A94_YCATS_W 0.708
A97_YCATS_W 0.704
A103_YCAT_A 0.711
A128_STRS 0.814
A147_ClsSup 0.830
131
Table 32. Continued
A151_ClsSup 0.781
A152_ClsSup 0.795
A156_ClsSup 0.817
Punishment A48_Friend 0.824
A50_Friend 0.693
A51_Friend 0.752
A104_YCAT_C 0.708
A106_YCAT_C 0.789
A107_YCAT_C 0.753
A108_YCAT_C 0.739
A111_YCAT_C 0.669
A112_YCAT_C 0.832
A122_STRS 0.843
A149_ClsSup 0.862
Autonomy-Support A98_YCAT_A 0.715
A101_YCAT_A 0.623
A159_MePart 0.652
A160_MePart 0.820
A164_MePart 0.726
A165_MePart 0.769
A166_MePart 0.712
Perceptions of Teacher Aware Understand 0.895
Caring 0.981
Perceptions of Self Affect Experience 0.909
Closeness 0.974
Interpersonal Scripts Emotional Reliance 0.893
Social Support Seeking 0.911
Internalization A171_Add_In 0.675
A172_Add_In 0.696
A173_Add_In 0.693
A174_Add_In 0.839
Behavioral Engagement A175_Add_En 0.702
A176_Add_En 0.864
Emotional Engagement A177_Add_En 0.881
A178_Add_En 0.906
A179_Add_En 0.871
Achievement Course Grade 1.000
Path Coefficients
Involvement --> Perceptions of Teacher 0.919
Involvement --> Perceptions of Self 0.931
Punishment --> Perceptions of Teacher 0.078
Punishment --> Perceptions of Self 0.060
Autonomy-Support --> Perceptions of Teacher 0.027
Autonomy-Support --> Perceptions of Self 0.031
Perceptions of Teacher --> Interpersonal Scripts -15.301
Perceptions of Self --> Interpersonal Scripts 16.306
Interpersonal Scripts --> Internalization 0.847
132
Table 32. Continued
Interpersonal Scripts --> Behavioral Engagement 0.573
Interpersonal Scripts --> Emotional Engagement 0.858
Internalization --> Achievement -17.737
Behavioral Engagement --> Achievement 2.349
Emotional Engagement --> Achievement 16.478
With respect to hypotheses five, six, and seven, invariance was not examined for
participant gender, teacher gender, or course topic groupings. Several path coefficients in
(Joreskog, 1999; Kline, 2005) and 2) a strong suppressor structure creating negative
suppression effects in the model (Kline, 2005). For example, the path coefficients from
multicollinearity issue. The β = .919 and β = .931 paths from involvement suggest
predictive importance and the β = .078, β = .060, β = .027 and β = .031 paths from
punishment and autonomy support respectively suggest those variables are unimportant.
When considered alongside of the subscale correlations displayed in Table 32, there is an
the paths from perceptions of teacher (β = -15.301) and perceptions of self (β = 16.306)
coefficients (Joreskog, 1999) and negative suppression due to the reversal of effect
directionality indicated by the coefficient (Kline, 2005). Paths from the outcome factors
133
that the path coefficient values in the model are of little use except to demonstrate the
scores for the predictor variables (involvement, punishment, and autonomy support) and
engagement) were calculated from the associated measured items in the final full model
as had already been done for the relational schema component factors
Table 33. Bivariate correlations among subscale scores for predictor factors, relational
schema factors, and outcome factors in modified full structural model
Subscale 1 2 3 4 5 6 7 8 9 10 11 12
1. Involvement 1 .78* .54* .89* .96* .95* .89* .90* .86* .74* .54* .82*
2. Punishment 1 .35* .62* .81* .76* .81* .65* .61* .57* .50* .61*
* * * * * * * *
3. Autonomy-Support 1 .55 .54 .55 .47 .58 .56 .39 .23 .49*
4. Aware/Understand 1 .87* .90* .79* .92* .90* .66* .48* .74*
5. Caring 1 .95* .89* .88* .84* .69* .51* .77*
6. Closeness 1 .90* .91* .89* .71* .50* .79*
* * * *
7. Affective Experience 1 .81 .79 .67 .47 .79*
8. Social Support Seeking 1 .94* .64* .45* .73*
9. Emotional Reliance 1 .65* .45* .73*
10. Internalization 1 .62* .80*
11. Behavioral Engagement 1 .60*
12. Emotional Engagement 1
* Correlation is significant at the 0.01 level (2-tailed).
and emotional reliance). Next, bivariate correlations were calculated for all of these
variables. As displayed in Table 33, all of these variables are significantly intercorrelated
with numerous correlations surpassing the r > .85 criterion suggested by Kline (2005) as
134
as the dependent variable to acquire a report of the collinearity diagnostic statistics. The
resulting variance inflation factor (VIF) statistics are reported in Table 34. It has been
suggested that a VIF of 5.0 or greater is something to be concerned about and VIF of 10.0
and spread across various levels of the model. Dissecting, deleting, recombining, and
Instead, exploratory factor analysis as presented in the post hoc analysis section generates
model, exploratory factor analysis was utilized to discover underlying empirical structure
in the data. Principal components analysis method of extraction was used because the
end goal was to reduce the set of items to a small number of factors that might be used to
135
create composite scores for use in subsequent analysis. Because avoiding
rotation, developed by Kaiser (1958), was used. The scree plot (Figure 35) was
examined to see if there was an obvious point at which the last substantial drop in the
magnitude of the eigenvalues would lead to a clear decision about the number of factors
Figure 35. Forward portion of the principal components analysis scree plot
to extract, which is known as the scree test (Cattell, 1966). The most substantial drop
was after the first factor, suggesting a possible best model might be the single factor
solution. However, the second and third factors were retained for extraction as they also
well.
Accordingly, principal components analysis was run with all 170 items from the
C-STRI that came from the 14 original instruments using VARIMAX rotation and
136
extracting 3 factors. The additional 10 items that were added to improve the
included as they were never thought to be measures of the student teacher relationship per
se. To reduce the number of items and clean up the 3-factor extraction results, item
elimination began with deletion of all items that did not have a primary factor loading of
.6 or greater. This step resulted in the deletion of 43 items (see Appendix J for factor
loadings). The principle components analysis was run again with the retained items and
the second step in item elimination was performed. This eliminated all items that had a <
.2 gap between their primary factor loading and any cross-loading. This step resulted in
the deletion of 31 more items (see Appendix K for factor loadings). The analysis was run
again. The third and final item elimination strategy removed all items with a cross-
loading of .3 or above. This step resulted in 58 more items being deleted (see Appendix
L for factor loadings). The results of this process are displayed in Table 35. Although
the retained items formed three distinct factors empirically, examination of their content
suggested an obvious methods effect with the first factor containing positive relationship
items and the second factor containing negatively worded items. The third factor is not a
measure of the relationship at all; rather, it contains items related a specific teaching
strategy of providing students with opportunities for choice and decision making in the
classroom. In other words, the results provide support for a global relationship quality
construct hypothesized in this study. While there are numerous items retained in the 3-
factor model (Figure 34) that could be deleted for redundancy, it would not be
137
Table 35. Factor loadings of items retained for the three factor solution after items
deleted for issues of cross-loading
Factor
Item Code Wording 1 2 3
A135_NRS_I In my relationship with this teacher, I felt close to them. .804 .297 .249
A148_ClsSup This teacher’s was like family to me. .803 .137 .194
A126_STRS If upset, I would seek comfort from this teacher. .790 .224 .208
A18_IPPA_C I could count on this teacher when I needed to get something off my chest. .777 .277 .249
A16_IPPA_c I told this teacher about my problems and troubles. .734 .026 .210
A29_Inv_Att This teacher knew a lot about me. .733 .243 .181
A13_IPPA_C This teacher could tell when I was upset about something. .711 .245 .148
A136_NRS_I In my relationship with this teacher, I felt attached to them. .706 .181 .258
A117_STRS I allowed this teacher to connect/relate with me. .694 .269 .286
A15_IPPA_C This teacher helped me to understand myself better. .692 .287 .230
A17_IPPA_C This teacher helped me to talk about my difficulties. .689 .231 .255
A11_IPPA_C I liked to get this teacher’s point of view on things I was concerned about. .677 .291 .281
A125_STRS This teacher shared an affectionate, warm relationship with me. .653 .295 .288
A121_STRS I tried to get support and encouragement from this teacher. .652 .079 .084
If I were having trouble understanding a subject at school, I would talk it
A141_Sch_Ut .625 .295 .221
over with this teacher.
A149_ClsSup This teacher was mean to me. .253 .831 .122
A122_STRS I saw this teacher as a source of punishment and criticism. .254 .793 .123
A48_Friend This teacher told me publicly that I might get a bad grade in that subject. .178 .781 .145
A112_YCAT_C This teacher was mean. .292 .771 .159
A106_YCAT_C This teacher got angry with me. .150 .770 .129
This teacher seemed to expect me to do shoddy work or make stupid
A51_Friend .180 .745 .115
mistakes.
A107_YCAT_C This teacher told me that I was doing something wrong a lot. .149 .736 .128
A104_YCAT_C This teacher told me I was going to get in trouble a lot. .062 .730 .084
A75_PSSM It was hard for people like me to be accepted by this teacher. .288 .709 .138
A108_YCAT_C This teacher told me that I didn’t try hard enough. .154 .696 .109
A111_YCAT_C This teacher told me I did not listen. .071 .679 .066
This teacher seemed pessimistic about my ability to be self-disciplining
A50_Friend .144 .678 .129
and responsible for my own behavior.
This teacher threatened to give me more work, tests, or to lower my grade
A53_Friend .024 .658 .028
to control my behavior.
A109_YCAT_C This teacher told me to stop doing work I liked doing. .162 .645 .091
Talking over my problems with this teacher made me feel ashamed or
A20_IPPA_A .286 .637 .088
foolish.
A160_MePart This teacher let me have a say in deciding what went on in class. .254 .121 .765
A98_YCAT_A This teacher let me choose work that I wanted to do. .265 .131 .736
A157_MePart This teacher let me help decide what the rules were going to be. .195 .031 .736
A166_MePart This teacher let me choose what I would work on. .280 .154 .717
A164_MePart This teacher included me when planning what we did in class. .248 .037 .676
A101_YCAT_A This teacher let me do different activities in class. .175 .155 .645
A159_MePart This teacher let me get a rule changed if I thought it was unfair. .276 .173 .619
A163_MePart In this class the teacher was the only one who decided on the rules. .115 .170 .611
Factor Correlations
Factor 1 1.00 .533 .581
Factor 2 1.00 .360
Factor 3 1.00
illustration of distinct relationship dimensions. Only 22%, 38 out of 170 items were
retained using the item deletion process outlined above, which eliminated items based on
138
how indistinctive they were. This further illustrates how extremely inter-correlated the
items were, why the hypothesized latent variables displayed extreme multicollinearity,
and gave additional support for the potential that a single factor solution might be best.
As previously mentioned, the scree plot in Figure 35 provided additional evidence that a
single-factor extraction might produce the best results. Therefore, the final step in this
iterative process of exploratory, post hoc data analysis was to run the principle
components analysis once more, this time with the restriction of extracting one factor.
Table 36. Factor loadings of all items for the single-factor solution
Item Wording Loading
A130_NRS_A In my relationship with this teacher, I felt supported. .893
A54_Rel When I was with this teacher I felt accepted. .889
A132_NRS_A In my relationship with this teacher, I felt listened to. .885
A131_NRS_A In my relationship with this teacher, I felt understood. .884
A133_NRS_A In my relationship with this teacher, I felt valued. .882
A76_CLM_SS This teacher really cared about me. .875
A27_Inv_Aff This teacher really cared about me. .868
A153_ClsSup This teacher cared about me. .868
A8_IPPA_T This teacher understood me. .868
A167_RAPS_E When I was with this teacher, I felt good. .864
A79_CLM_SS This teacher cared about my feelings. .863
A170_RAPS_E When I was with this teacher, I felt happy. .863
A10_IPPA_T I trusted this teacher. .862
A35_Inv_Dep I could count on this teacher to be there for me. .857
A74_PSSM This teacher liked me the way I was. .856
A49_Friend This teacher was warm and supportive. .854
A58_PSSM I felt like a real part of this teacher’s class. .854
A34_Inv_Dep This teacher was always there for me. .852
A4_IPPA_T This teacher accepted me as I was. .852
A1_IPPA_T This teacher respected my feelings. .849
A114_STRS I trusted this teacher. .848
A73_PSSM I feel proud of knowing this teacher. .847
A3_IPPA_T I wish I had a different teacher. .846
A84_BNS_Rel When I was with this teacher, I felt loved and cared about. .846
A119_STRS This teacher was able to comfort me or cheer me up. .843
A60_PSSM This teacher took my opinions seriously. .842
If I were feeling excited about something happening in my life, I would have been
A42_ERQ .841
willing to turn to this teacher.
A36_Inv_Dep I could rely on this teacher to be there when I needed him/her. .840
A154_ClsSup This teacher treated me with respect. .839
A37_Inv_Dep This teacher was never there for me. .839
139
Table 36. Continued
A70_PSSM This teacher respected me. .838
A57_Rel When I was with this teacher I felt unimportant. .837
A2_IPPA_T I feel this teacher did a good job as my teacher. .833
A63_PSSM I could talk to this teacher if I had a problem. .831
A169_RAPS_E When I was with this teacher, I felt unhappy. .830
A92_YCATS_W This teacher made the class fun. .827
If I were feeling proud of my accomplishments, I would have been willing to turn
A46_ERQ .826
to this teacher.
A72_PSSM I wished I was in a different teacher’s class. .825
A82_CLM_AS This teacher liked to help me learn. .825
A67_PSSM I was treated with respect by this teacher. .823
A135_NRS_I In my relationship with this teacher, I felt close to them. .823
A88_YCATS_W This teacher liked me. .823
A6_IPPA_T When we discussed things, this teacher cared about my point of view. .822
A144_Emul I would feel good if someone said I was a lot like this teacher. .822
A56_Rel When I was with this teacher I felt ignored. .821
A26_Inv_Aff This teacher liked me. .821
A142_Sch_Ut I could usually rely on this teacher when I had problems at my school. .820
A147_ClsSup When I did well, this teacher made me feel good. .819
A120_STRS I avoided contact with this teacher. .818
A31_Inv_Att This teacher didn’t understand me. .818
If I were feeling very bad about myself and needed a boost, I would have been
A43_ERQ .818
willing to turn to this teacher.
A124_STRS I felt safe and secure with this teacher. .817
A7_IPPA_T This teacher trusted my judgment. .817
A69_PSSM I could really be myself around this teacher. .816
If I were feeling anxious or scared about something, I would have been willing to
A41_ERQ .815
turn to this teacher.
My relationship with this teacher became more positive throughout the school
A128_STRS .813
year.
A55_Rel When I was with this teacher I felt like someone special. .811
A44_ERQ If I were feeling overwhelmed, I would have been willing to turn to this teacher. .810
A138_NRS_I In my relationship with this teacher, I felt close-knit. .810
A155_ClsSup This teacher helped me, even if he/she was busy. .809
A90_YCATS_W This teacher said nice things about my work. .809
A9_IPPA_T When I was angry about something, this teacher tried to be understanding. .809
A64_PSSM This teacher was friendly to me. .809
A91_YCATS_W This teacher helped me when I didn’t understand. .806
A89_YCATS_W This teacher was my friend. .804
A59_PSSM This teacher noticed when I was good at something. .799
A156_ClsSup This teacher helped me learn. .799
A115_STRS This teacher was in tune with what I was feeling. .798
A137_NRS_I In my relationship with this teacher, I felt bonded to them. .797
A80_CLM_AS This teacher cared about how much I learned. .793
A61_PSSM This teacher was interested in me. .792
A145_Emul I wanted to be like this teacher. .791
A18_IPPA_C I could count on this teacher when I needed to get something off my chest. .786
A33_Inv_DR This teacher talked with me. .783
A152_ClsSup This teacher worked with me to solve problems. .783
A28_Inv_Aff This teacher didn’t seem to enjoy having me in his/her class. .779
A30_Inv_Att This teacher knew me well. .777
140
Table 36. Continued
A23_IPPA_A I felt angry with this teacher. .774
If I were feeling frustrated or angry, I would have been willing to turn to this
A45_ERQ .771
teacher.
A168_RAPS_E When I was with this teacher, I felt mad. .771
A151_ClsSup When I had trouble with my schoolwork, this teacher tried to help. .769
A65_PSSM This teacher was not interested in people like me. .766
A139_NRS_I In my relationship with this teacher, I felt like a friend. .766
If I were feeling alone or depressed, I would have been willing to turn to this
A40_ERQ .766
teacher.
My relationship with this teacher became more negative throughout the school
A129_STRS .766
year.
A81_CLM_AS This teacher liked to see my work. .761
A93_YCATS_W This teacher answered my questions. .755
A24_IPPA_A I didn’t get much attention from this teacher. .753
A143_Emul I tried to model myself after this teacher. .751
A19_IPPA_C If this teacher knew something was bothering me, he/she asked me about it. .750
A112_YCAT_C This teacher was mean. .748
A149_ClsSup This teacher was mean to me. .747
A116_STRS I was wary of this teacher. .747
A146_Emul As I get older, I’ll probably be a lot like this teacher. .747
A11_IPPA_C I liked to get this teacher’s point of view on things I was concerned about. .744
A134_NRS_A In my relationship with this teacher, I felt safe. .744
A32_Inv_DR This teacher spent time with me. .743
A85_BNS_Rel When I was with this teacher, I often felt a lot of distance in our relationship. .741
A126_STRS If upset, I would seek comfort from this teacher. .740
A117_STRS I allowed this teacher to connect/relate with me. .740
A15_IPPA_C This teacher helped me to understand myself better. .737
A66_PSSM This teacher made sure I was included in lots of activities in class. .733
A78_CLM_SS This teacher liked me as much as other students. .732
A12_IPPA_C I felt it was no use letting my feelings show around this teacher. .730
A83_CLM_AS This teacher wanted me to do my best in schoolwork. .720
A125_STRS This teacher shared an affectionate, warm relationship with me. .719
A122_STRS I saw this teacher as a source of punishment and criticism. .716
A150_ClsSup This teacher just looked out for his/herself. .715
A29_Inv_Att This teacher knew a lot about me. .712
A94_YCATS_W This teacher told good stories. .711
A75_PSSM It was hard for people like me to be accepted by this teacher. .707
A103_YCAT_A This teacher told me I was smart. .706
A97_YCATS_W This teacher smiled a lot. .706
A140_Sch_Ut If I had a problem with my school work I would share it with this teacher. .704
A21_IPPA_A I got upset easily around this teacher. .702
A13_IPPA_C This teacher could tell when I was upset about something. .702
A62_PSSM Sometimes I felt as if I didn’t belong in this teacher’s class. .700
A17_IPPA_C This teacher helped me to talk about my difficulties. .700
A25_IPPA_A This teacher didn’t understand what I was going through in those days. .692
If I were having trouble understanding a subject at school, I would talk it over with
A141_Sch_Ut .692
this teacher.
A148_ClsSup This teacher’s was like family to me. .682
A136_NRS_I In my relationship with this teacher, I felt attached to them. .671
A48_Friend This teacher told me publicly that I might get a bad grade in that subject. .667
A68_PSSM I felt very different from this teacher. .666
141
Table 36. Continued
A71_PSSM This teacher knew I could do good work. .665
A38_Inv_Dep I couldn’t depend on this teacher for important things. .640
A77_CLM_SS This teacher thought it was important to be my friend. .638
A51_Friend This teacher expected me to make stupid mistakes. .635
A87_YCATS_W This teacher liked my family. .633
A102_YCAT_A This teacher did activities with me. .628
A20_IPPA_A Talking over my problems with this teacher made me feel ashamed or foolish. .628
A95_YCATS_W This teacher remembered special days for me. .627
A106_YCAT_C This teacher got angry with me. .619
A107_YCAT_C This teacher told me that I was doing something wrong a lot. .610
A86_BNS_Rel When I was with this teacher, I felt a lot of closeness and intimacy. .601
A100_YCAT_A This teacher let me do activities I wanted to do. .589
A123_STRS I participated deeply in this teacher’s class. .588
A108_YCAT_C This teacher told me that I didn’t try hard enough. .580
This teacher seemed pessimistic about my ability to be self-disciplining and
A50_Friend .574
responsible for my own behavior.
A158_MePart This teacher let me do things that I wanted to do. .572
A162_MePart This teacher let me do things my own way. .571
A16_IPPA_c I told this teacher about my problems and troubles. .561
A113_YCAT_C This teacher had too many rules for our class. .556
A39_Inv_Dep I couldn’t count on this teacher when I needed him/her. .545
A109_YCAT_C This teacher told me to stop doing work I liked doing. .543
A104_YCAT_C This teacher told me I was going to get in trouble a lot. .528
A121_STRS I tried to get support and encouragement from this teacher. .524
A166_MePart This teacher let me choose what I would work on. .500
A111_YCAT_C This teacher told me I did not listen. .493
A159_MePart This teacher let me get a rule changed if I thought it was unfair. .488
A98_YCAT_A This teacher let me choose work that I wanted to do. .481
A160_MePart This teacher let me have a say in deciding what went on in class. .480
A22_IPPA_A I get upset a lot more than this teacher knew about. .478
A14_IPPA_C This teacher had his/her own problems, so I didn’t bother him/her with mine. .473
A96_YCATS_W This teacher chose me to be a special helper. .471
A5_IPPA_T This teacher expected too much of me. .468
A161_MePart This teacher let me help plan what we would do in class. .453
A165_MePart This teacher asked me to help decide what the class should do. .449
A118_STRS I tried to get help, recognition, and support from this teacher. .424
This teacher threatened to give me more work, tests, or to lower my grade to
A53_Friend .422
control my behavior.
A101_YCAT_A This teacher let me do different activities in class. .421
A164_MePart This teacher included me when planning what we did in class. .402
A157_MePart This teacher let me help decide what the rules were going to be. .374
A99_YCAT_A This teacher let me choose where I wanted to sit. .368
A127_STRS I felt upset when separated from this teacher. .360
A163_MePart In this class the teacher was the only one who decided on the rules. .356
A105_YCAT_C This teacher told me to do work that was too hard for me. .308
A110_YCAT_C This teacher told me to do work I didn’t want to do. .307
A47_Friend This teacher criticized me for turning work in late or failing to turn in assignments. .304
A52_Friend This teacher used sarcasm. -.052
142
The single-factor loadings offer strong support for a general, global relationship
quality construct. Aside from one outlier item “this teacher used sarcasm” (λ = -.052) the
lowest loading was λ = .304, which is above the cutoff specified previously for an item to
be acceptably distinct from the factor. On the other hand, 65 items loaded λ > .800, 118
items loaded λ > .700, and 135 items loaded λ > .600, which was the cutoff specified
previously for acceptance as a primary factor loading. That is, nearly 80% of the items
loaded strongly in the single-factor model and 99% of the items resulted in λ > .300,
which was used as a cutoff for indicating “salient” factor loadings in a Psychological
Methods journal report on the use of exploratory factor analysis in psychological research
143
CHAPTER V. DISCUSSION
Several interesting findings resulted from this study. The hypothesized model,
representation of their relationship with a teacher, was not supported by the data. The
study did, however, provided a first look into the comparative utility of 14 different
model fit statistics, and then be successfully recombined into a series of other
measurement models that fit the data as well. Although not in the expected fashion, the
study also afforded empirical evaluation as to which of the 170 items from the 14 original
scales most closely measure the core of student teacher relationship quality and which
items do not. The study exemplified the dangers of negative item wording. Finally the
All of the original 14 scales were examined using confirmatory factor analysis to
check model fit and the results were presented in Table 3. Based on the data collected for
this study, quality of fit varied substantially for the original scales. The Basic Need
Satisfaction in Relationships scale had only one factor with three items so the model was
“just identified” (Kline, 2005) and thus its CFA results are not to be trusted. From the
data collected for this study, five scales produced unacceptable model fit statistics.
Specifically, the Research Assessment Package for Schools generated RMSEA = .326,
144
Teacher as a Social Context generated RMSEA = .104, Sense of Relatedness generated
RMSEA = .162, Student Teacher Relationship Scale generated RMSEA = .128 & CFI =
.849, and the Young Children’s Appraisals of Teacher Support generated CFI = .860.
The remaining 8 scales demonstrated acceptable fit RMSEA = .037-.090 and CFI = .912-
.995. A few of these original scales have the results of exploratory factor analyses
reported in the original publication of the scales, none of them reported measurement
model fit statistics from confirmatory factor analyses. Therefore, the confirmatory factor
analysis results found in this project could not be compared to original model fit
statistics. However, all of the original studies reported internal reliability coefficient
alphas. Table 37 shows the original coefficient alphas and the coefficient alphas found in
this study side-by-side for each of the 14 scales and respective subscales. In comparison
to internal reliability data presented in the original publications for each of the 14 scales,
data from this study suggest slightly more internal consistency; however the patterns of
reliability coefficients match fairly well. All scales and subscales demonstrated
results (βeta & R2) from Table 18 (internalization), Table 19 (engagement), and Table 20
(achievement) were condensed into a single table. Bivariate correlations were added as
well due to concerns about suppressor effects in the βeta coefficients. Table 38 shows the
and Beta coefficients revealed suppressor effects in the scales that were divided into
subscales. The combination of these statistics and the post hoc exploratory factor
analysis call the proposed dimensionality of these original scales into question. More
145
Table 37. Comparison of internal reliability from original publication and this study for
all 14 original scales
Scale Subscale Items Original α This Study α
Social Support 4 .80 .86
Classroom Life Measure (N = 91)
Academic Support 4 .78 .90
Student Classroom Environment
Friendly/Fair 7 .70 - .75 .76
Measure (N = 1788)
Developmental Studies Center Classroom Supportiveness 10 .82 .94
Student Questionnaire (N = 5143) Meaningful Participation 10 .80 .91
Trust 10 .91 .95
Inventory of Parent & Peer
Communication 9 .91 .92
Attachments (N = 179)
Alienation 6 .86 .86
School Utilization & Teacher School Utilization 3 .66 .83
Emulation (N = 606) Teacher Emulation 4 .84 .94
Research Assessment Package
Emotional Security 4 .71 .92
for Schools (N = 2429)
Affection 3 .71 .88
Teacher as Social Context (N = 500) Attunement 3 .54 .86
Dedication of Resources 2 NA .77
Dependability 6 .72 .90
Sense of Relatedness (N = 641) NA 4 .79 .91
Emotional Reliance
NA 7 .91 - .97 .96
Questionnaire (N = 195)
Basic Need Satisfaction in
Relatedness 3 .90 .79
Relationships (N = 152)
Need for Relatedness Scale Acceptance 5 .89 .94
(N = 265) Intimacy 5 .91 .93
Student-Teacher Relationship Scale
NA 16 .85 .94
(N = 72)
Warmth 11 .75 .92
Young Children’s Appraisals
Autonomy 6 .67 .79
of Teacher Support (N = 364)
Conflict 10 .75 .88
Psychological Sense of School
Belongingness 18 .87 .97
Membership (N = 454)
will be said about support for a single global relationship quality construct later. Overall,
the data in Table 38 suggest that all of the relationship measures have predictive validity
in relation to internalization, engagement, and achievement. When all three outcomes are
considered together, the Student Classroom Environment Measure, School Utilization &
146
Teacher Emulation, the Emotional Reliance Questionnaire, and the Basic Needs
147
Table 38. Predictive validity comparison of 14 original scales
Internalization Engagement Achievement
Scale Subscale r β R2 r β R2 r β R2
Student Classroom
Friendly/Fair .52** .52*** .27*** .58** .58*** .34*** .45** .45*** .20***
Environment Measure
Developmental Studies Center Classroom Supportiveness .71** .69*** .78** .74*** .45** .38***
.50*** .61*** .22***
Student Questionnaire Meaningful Participation .39** .03 .47** .08** .34** .14**
School Utilization & Teacher School Utilization .63** .31*** .70** .34*** .41** .30*** .17***
.48*** .58***
Emulation Teacher Emulation .66** .43*** .73** .47*** .37** .14**
Sense of Relatedness NA .68** .68*** .46*** .75** .75*** .57*** .45** .45*** .20***
Emotional Reliance
NA .64** .64*** .41*** .71** .71*** .50*** .40** .40*** .16***
Questionnaire
148
Table 38. Continued
Basic Need Satisfaction in
Relatedness .56** .56*** .32*** .65** .65*** .43*** .38** .38*** .15***
Relationships
Student-Teacher Relationship
NA .71** .71*** .51*** .79** .79*** .62*** .46** .46*** .21***
Scale
149
Multiple Models Can Fit the Data
like the one used in this project can be problematic. One might even say that this project
serves as a prime example of how the use of model fit to judge the quality of a model can
be misleading in some cases. Even if a model is consistent with data, a researcher cannot
In this study, each of the measurement models, with only slight modifications
required, came to “fit” the data acceptably well. A researcher’s experience as this
happens is typically of relief and satisfaction. Moreover, this feeling was experienced
tenfold as the full model underwent its final modification and it too “fit” the data.
However, the path coefficients, multicollinearity analyses, and exploratory post hoc data
analyses together remind us of the fact that even when a model fits, the researcher should
always explore alternative models. Perhaps Norman Cliff (1983, 118, italics in original)
well to remember that models other than the one that “fits” will fit the data equally well.
Indeed, the very form of the equations underlying LISREL guarantee that in virtually
every application there are an infinity of models that will fit the data equally well.”
questionnaire has been encouraged by experts in educational testing for a long time (e.g.,
Anastasi, 1982; Cronbach 1950; Spector, 1992). The rationale is that reversed wording
helps to avoid participant bias such as response acquiescence or to add mental speed
150
bumps that slow down cognitive processing to improve the thoughtfulness of responses
studies of self-esteem scales have produced evidence that including both items worded
negatively and items worded positively can thwart accurate analysis and interpretation of
survey data by creating systematic measurement error (Horan, DiStefano, & Motl, 2003;
DiStefano & Motl, 2006; 2009; Marsh, 1996). Similar evidence is demonstrated in the
data collected for this study as displayed in Table 35. The student teacher relationship
between factors 1 and 2 is clearly the wording of the items as positive statements or
One could argue that the two factors represent positive aspects of the relationship
(warmth) and negative aspects of relationship (conflict) and that the items in factor 3
represent a third factor (autonomy). This is the model proposed by Mantzicopoulos and
Neuharth-Pritchett (2003). However, the factor loadings for the single factor model
(Table 36) cast strong doubt on the distinction between the warmth and conflict
dimensions. The items loading on factor 3 in Table 35 do include items from the Y-
CATS instrument’s autonomy subscale and similar items about opportunities for students
to experience choices in the classroom from the meaningful participation subscale of the
Developmental Studies Center Student Questionnaire (DSC, 2002) and they do create a
distinct factor; yet, these items represent a teaching strategy that may or may not have an
151
influence on the relationship. Further, there is much more to autonomy-support than
Other researchers have wrestled with this same issue when measuring attachment
using the Inventory of Parent and Peer Attachment (Armsden & Greenberg, 1987). The
initial report stated that based on the eigenvalue criterion of greater than 1, they extracted
subscales (r between .70 and .76), the authors utilized a combined score for overall
attachment instead of the individual factor scores for the analyses. Johnson, Ketring, and
Abshire (2003) ran a confirmatory factor analysis on the IPPA and found that it did not fit
(χ2 = 735.32, p < .001; CFI = .73; RMSEA = .09). When they ran the data using
exploratory factor analysis, two factors emerged. The first factor they found relates to
trust/avoidance and the items seem to represent the general relationship quality construct.
On the other hand, the second dimension they found had 6 items loading λ = .594 to .401
(N = 1059) Italian adolescents, Pace et al. (2011) found acceptable fit of a uni-factorial
solutions of the IPPA items, which corroborates the evidence in this dissertation that
Taking another angle, we ought to reflect again upon the performance of the
original instruments. The strength of the suppressor effects displayed among all of the
scales that were originally proposed to be multidimensional in Table 38 suggests that the
predictive power within each of the subscales is shared among them. In fact, the
152
subscales may not truly represent distinct factors at all. Instead, they may all be artifacts
to examine the bivariate correlations among the subscales. For the Student Classroom
Environment Measure, the Social Support and Academic Support subscales are highly
correlated (r = .82). A correlation of this magnitude provides a strong case for the single
factor model. The Developmental Studies Center Student Questionnaire, the Classroom
.53), which suggests that they contain some distinct elements but nonetheless share a
large amount of variance. Remember, several of the Meaningful Participation items were
among the items that loaded on the third factor in the exploratory factor analysis where
three factors were extracted. Those items, measuring the allowance of student choices as
a classroom strategy, are likely responsible for lowering the bivariate correlation. Similar
to Armsden and Greenberg’s (1987) original findings, the Inventory of Parent and Peer
Attachment subscales were highly correlated (r = .73 - .85). School Utilization and
Teacher Emulation were also highly correlated (r = .76), as were the subscales of the
Teacher as a Social Context scale (r = .75 - .84), and the Acceptance and Intimacy
subscales of the Need for Relatedness Scale (r = .82). Finally, the Young Children’s
Appraisals of Teacher Support scale produced subscale correlations of r = .51 - .76 with
the lower correlations being those that relate to the Autonomy subscale, which includes
the other items that made up the third factor in the three-factor exploratory factor analysis
model.
The fact that half of the original 14 scales measured the quality of the relationship
as a single factor should not be left unmentioned. All seven of the single factor scales
153
produced reasonable predictive validity as reported in Table 38 and one of them, the
Psychological Sense of School Membership, arguably performed the best overall. All of
this evidence, along with the importance of selecting the most parsimonious theory,
Limitations
This study has several limitations. First, the data were collected from college
students’ retrospective reports of relationships with high school teachers. The reasoning
for this is that college students are a convenient population to study; however, their
relationships with current college professors may have been under developed depending
on the time of participation in the study and because relationships with college professors
are often less developed due to the format of collegiate instruction, especially in the large
lecture classes of freshman and sophomores. In most cases, the relationships participants
were asked to recall occurred between 1-2 years ago, but in some cases up to 5 years ago.
remember their relationships with several high school teachers. That said, the fact that
they are all university students separates them from being a representative high-school
student population as many high school students do not attend a university. Therefore
this sample may have been biased in terms of higher than average achievement.
inaccurate or subject to cognitive bias. Ideally, predictor variables and outcome variables
as they are happening rather than memories skewed by known systematic errors of
154
cognitive bias such as confirmation bias. However, it should be noted that the purpose of
this study was to explore the underlying structure of students’ internal representations of
That said, an argument could still be made that the passing of time causes a
reduction in the level of detail contained in students’ relation schemas. That is, there may
time passes that representation may fade into a generalized, relatively positive or
Second, also related to the study being based on self-report, there is no way to
know if participants put effort into making distinct responses to the massive variety of
items. It is possible that the huge number of items led participant to speed up the process
of filling out the survey by relying on a global assessment of whether the relationship was
positive or negative rather than fully reading each item or putting much thought into their
responses. There was also the potential problem of reactivity where participants may
have begun to think about what is socially desirable or what the researcher was looking
the Midwestern university student population from which the convenience sample came,
the sample was not very diverse in terms of ethnicity, age, or socioeconomic status.
relationships under examination to those from the final high school years. In other
words, it is possible that these student-teacher relationships do not reflect what is most
relevant in students’ relationships with teachers before and/or after the high school years.
155
General Conclusions and Next Steps
On one hand, the data from this study may suggest that elements of self-
indistinct as relational schema components. That is, judgements about the teacher,
judgements about one’s self in relation to the teacher, and expectations about interactions
with the teacher may all be dependent on the student’s global assessment of the quality of
their relationship with that teacher. On the other hand, the data may suggest that aspects
coherent and so highly interconnected that they failed to reveal any empirical distinction
working models, are well established, as reviewed in the literature that led to the
hypotheses formulated for this study. Unfortunately, empirical tests of those theoretical
propositions are lacking and the data collected in this study does little to support them;
rather, these data suggest that such complexities may be unnecessary when examining the
outcomes.
variety of social cognitive methodology are still required. For example, a distinction
between relational schema components may be better assessed using implicit priming
156
techniques as opposed to retrospective self-reports. In the adult attachment literature,
researches have had some success using implicit methods to examine cognitive processes
underlying working models, however, direct evidence of structure and function does not
As usual, we are left with more questions than answers. Are student-teacher
relational schemas multidimensional? Are student’s relational schemas the best construct
to examine when measuring relationship quality with their teachers? Are relational
an existing relationship and more general with the passage of time following an end to
the relationship? Further research is required to address questions like these and more.
157
APPENDIX A. ORIGINAL 14 INSTRUMENTS, SUBSCALES, AND ITEMS
Inventory of Parent & Peer Attachment – Revised Version (Armsden & Greenberg, 1987)
1) My __ respects my feelings.
4) My __ accepts me as I am.
7) My __ trusts my judgment.
8) My __ understands me.
Communication (9 items).
12) I feel it is no use letting my feelings show around my __. (reverse score)
14) My __ has her own problems, so I don’t bother her with mine. (reverse score)
158
Alienation (6 items). (all reverse scored)
Teacher as Social Context (TASC; Belmont, Skinner, Wellborn, & Connell, 1992)
Affection (3 items).
28) My teacher doesn’t seem to enjoy having me in her class. (reverse score)
Attunement (3 items).
Dependability (6 items).
159
37) My teacher is never there for me. (reverse score)
No subscales (7 items).
40) If I were feeling alone or depressed, I would be willing to turn to (my) _____.
43) If I were feeling very bad about myself and needed a boost, I would be
45) If I were feeling frustrated or angry, I would be willing to turn to (my) _____.
(my) _____.
Teacher—unfair/unfriendly (7 items).
47) Students are criticized for turning math work in late or failing to turn in
48) The teacher says to some students or the class as a whole that they may get a
160
50) The teacher seems pessimistic about the ability of students to be self-
51) The teacher seems to expect some students to do shoddy work or make stupid
53) The teacher threatens to give more work, math tests, or to lower grades to
No subscales (4 items).
63) There’s at least one teacher or other adult in this school I can talk to if I have
a problem.
161
65) Teachers here are not interested in people like me. (reverse score)
68) I feel very different from most other students here. (reverse score)
162
Relatedness (3 items).
85) When I am with XXXXX, I often feel a lot of distance in our relationship.
(reverse score)
Pritchett, 2003)
Autonomy (6 items).
163
101) My teacher lets me do different activities in class.
118) This child seeks help, recognition, and support from me.
164
121) This child constantly needs reassurance from me.
122) This child sees me as a source of punishment and criticism. (reverse score)
128) My relationship with this child has become more positive during the school
year.
129) My relationship with this child has become more negative during the school
Acceptance (5 items).
Intimacy (5 items).
165
139) In my relationship with my teacher, I feel as a friend.
140) If I had a problem with my school work I would share it with my teacher.
Emulation (4 items).
144) I would feel good if someone said I was a lot like my teacher.
147) When someone in my class does well, everyone in the class feels good.
150) Students in my class just look out for themselves. (reverse score)
151) When I’m having trouble with my schoolwork, at least one of my classmates
153) The students in my class don’t really care about each other.
166
155) Students in my class help each other, even if they are not friends.
157) In my class the teacher and students decide together what the rules will be.
159) Students in my class can get a rule changed if they think it is unfair.
161) In my class the students get to help plan what they will do.
163) In my class the teacher is the only one who decides on the rules. (reverse
score)
164) In my class the teacher and students together plan what we will do.
165) The teacher in my class asks the students to help decide what the class
should do.
167
APPENDIX B. REVISED 14 INSTRUMENTS, SUBSCALES, AND ITEMS
Inventory of Parent & Peer Attachment – Revised Version (Armsden & Greenberg, 1987)
Communication (9 items).
11) I liked to get this teacher’s point of view on things I was concerned about.
12) I felt it was no use letting my feelings show around this teacher. (reverse
score)
13) This teacher could tell when I was upset about something.
14) This teacher had his/her own problems, so I didn’t bother him/her with mine.
(reverse score)
168
18) I could count on this teacher when I needed to get something off my chest.
19) If this teacher knew something was bothering me, he/she asked me about it.
20) Talking over my problems with this teacher made me feel ashamed or foolish.
22) I get upset a lot more than this teacher knew about.
25) This teacher didn’t understand what I was going through in those days.
Teacher as Social Context (TASC; Belmont, Skinner, Wellborn, & Connell, 1992)
Affection (3 items).
28) This teacher didn’t seem to enjoy having me in his/her class. (reverse score)
Attunement (3 items).
Dependability (6 items).
169
35) I could count on this teacher to be there for me.
37) This teacher was never there for me. (reverse score)
38) I couldn’t depend on this teacher for important things. (reverse score)
39) I couldn’t count on this teacher when I needed him/her. (reverse score)
No subscales (7 items).
40) If I were feeling alone or depressed, I would have been willing to turn to this
teacher.
41) If I were feeling anxious or scared about something, I would have been
42) If I were feeling excited about something happening in my life, I would have
43) If I were feeling very bad about myself and needed a boost, I would have
44) If I were feeling overwhelmed, I would have been willing to turn to this
teacher.
45) If I were feeling frustrated or angry, I would have been willing to turn to this
teacher.
Teacher—unfair/unfriendly (7 items).
170
47) This teacher criticized me for turning work in late or failing to turn in
48) This teacher told me publicly that I might get a bad grade in that subject.
(reverse score)
51) This teacher seemed to expect me to do shoddy work or make stupid mistakes.
(reverse score)
53) This teacher threatened to give me more work, tests, or to lower my grade to
No subscales (4 items).
55) When I was with this teacher I felt like someone special.
56) When I was with this teacher I felt ignored. (reverse score)
57) When I was with this teacher I felt unimportant. (reverse score)
171
61) This teacher was interested in me.
62) Sometimes I felt as if I didn’t belong in this teacher’s class. (reverse score)
65) This teacher was not interested in people like me. (reverse score)
66) This teacher made sure I was included in lots of activities in class.
75) It was hard for people like me to be accepted by this teacher. (reverse score)
172
81) This teacher liked to see my work.
Relatedness (3 items).
84) When I was with this teacher, I felt loved and cared about.
85) When I was with this teacher, I often felt a lot of distance in our relationship.
(reverse score)
86) When I was with this teacher, I felt a lot of closeness and intimacy.
Pritchett, 2003)
173
Autonomy (6 items).
105) This teacher told me to do work that was too hard for me.
107) This teacher told me that I was doing something wrong a lot.
113) This teacher had too many rules for our class.
174
117) I allowed this teacher to connect/relate with me.
118) I tried to get help, recognition, and support from this teacher.
122) I saw this teacher as a source of punishment and criticism. (reverse score)
128) My relationship with this teacher became more positive throughout the
school year.
129) My relationship with this teacher became more negative throughout the
Acceptance (5 items).
Intimacy (5 items).
175
135) In my relationship with this teacher, I felt close to them.
140) If I had a problem with my school work I would share it with this teacher.
142) I could usually rely on this teacher when I had problems at my school.
Emulation (4 items).
144) I would feel good if someone said I was a lot like this teacher.
150) This teacher just looked out for his/herself. (reverse score)
151) When I had trouble with my schoolwork, this teacher tried to help.
176
152) This teacher worked with me to solve problems.
157) This teacher let me help decide what the rules were going to be.
159) This teacher let me get a rule changed if I thought it was unfair.
160) This teacher let me have a say in deciding what went on in class.
163) In this class the teacher was the only one who decided on the rules. (reverse
score)
165) This teacher asked me to help decide what the class should do.
168) When I was with this teacher, I felt mad. (reverse score)
169) When I was with this teacher, I felt unhappy. (reverse score)
177
APPENDIX C. C-STRI ITEMS GROUPED BY PROPOSED MODEL FACTORS
66) This teacher made sure I was included in lots of activities in class.
105) This teacher told me to do work that was too hard for me. (reverse score)
151) When I had trouble with my schoolwork, this teacher tried to help.
47) This teacher criticized me for turning work in late or failing to turn in
assignments.
48) This teacher told me publicly that I might get a bad grade in that subject.
178
50) This teacher was pessimistic about my ability to be self-disciplining and
51) This teacher seemed to expect me to do shoddy work or make stupid mistakes.
53) This teacher threatened to give me more work, tests, or to lower my grade to
control my behavior.
107) This teacher told me that I was doing something wrong a lot.
Predictor: Involvement
24) I didn’t get much attention from this teacher. (reverse score)
179
37) This teacher was never there for me. (reverse score)
38) I couldn’t depend on this teacher for important things. (reverse score)
39) I couldn’t count on this teacher when I needed him/her. (reverse score)
56) When I was with this teacher I felt ignored. (reverse score)
128) My relationship with this teacher became more positive throughout the
school year.
129) My relationship with this teacher became more negative throughout the
150) This teacher just looked out for his/herself. (reverse score)
109) This teacher told me to stop doing work I liked doing. (reverse score)
110) This teacher told me to do work I didn’t want to do. (reverse score)
113) This teacher had too many rules for our class. (reverse score)
180
157) This teacher let me help decide what the rules were going to be.
159) This teacher let me get a rule changed if I thought it was unfair.
160) This teacher let me have a say in deciding what went on in class.
163) In this class the teacher was the only one who decided on the rules. (reverse
score)
165) This teacher asked me to help decide what the class should do.
28) This teacher didn’t seem to enjoy having me in his/her class. (reverse score)
65) This teacher was not interested in people like me. (reverse score)
181
74) This teacher liked me the way I was.
75) It was hard for people like me to be accepted by this teacher. (reverse score)
84) When I was with this teacher, I felt loved and cared about.
13) This teacher could tell when I was upset about something.
22) I get upset a lot more than this teacher knew about. (reverse score)
182
25) This teacher didn’t understand what I was going through in those days.
(reverse score)
55) When I was with this teacher I felt like someone special.
57) When I was with this teacher I felt unimportant. (reverse score)
85) When I was with this teacher, I often felt a lot of distance in our relationship.
(reverse score)
86) When I was with this teacher, I felt a lot of closeness and intimacy.
183
134) In my relationship with this teacher, I felt safe.
168) When I was with this teacher, I felt mad. (reverse score)
169) When I was with this teacher, I felt unhappy. (reverse score)
14) This teacher had his/her own problems, so I didn’t bother him/her with mine.
(reverse score)
18) I could count on this teacher when I needed to get something off my chest.
19) If this teacher knew something was bothering me, he/she asked me about it.
184
20) Talking over my problems with this teacher made me feel ashamed or foolish.
(reverse score)
142) I could usually rely on this teacher when I had problems at my school.
12) I felt it was no use letting my feelings show around this teacher. (reverse
score)
40) If I were feeling alone or depressed, I would have been willing to turn to this
teacher.
41) If I were feeling anxious or scared about something, I would have been
42) If I were feeling excited about something happening in my life, I would have
43) If I were feeling very bad about myself and needed a boost, I would have
44) If I were feeling overwhelmed, I would have been willing to turn to this
teacher.
45) If I were feeling frustrated or angry, I would have been willing to turn to this
teacher.
185
117) I allowed this teacher to connect/relate with me.
Outcome: Internalization
144) I would feel good if someone said I was a lot like this teacher.
*172) I felt that completing the assignments in this class was beneficial to me.
*174) Because of this teacher, I believe more in the purpose of learning the
Outcome: Engagement
11) I liked to get this teacher’s point of view on things I was concerned about.
62) Sometimes I felt as if I didn’t belong in this teacher’s class. (reverse score)
118) I tried to get help, recognition, and support from this teacher.
186
140) If I had a problem with my school work I would share it with this teacher.
* Indicates items that are not from the original 14 scales. They have been added to
187
APPENDIX D. TEXT FOR INVITATION PARTICIPATE EMAIL
188
APPENDIX E. ELECTRONIC INFORMED CONSENT DOCUMENT
189
APPENDIX F. INITIAL DEMOGRAPHIC QUESTIONS
190
APPENDIX G. END OF SURVEY THANK YOU NOTE
191
APPENDIX H. C-STRI STATEMENTS FOR COUSE SUBJECT ASSIGNMENT
192
193
APPENDIX I. SAMPLE OF C-STRI QUESTIONS FORMATTED AS A
*The first 3 items shown above were always delivered first; the remaining 179 items, 4 of
which are shown above, were delivered in an order randomized for each participant. See
Appendix C for a full list of the C-STRI Items.
194
APPENDIX J. PRINCIPAL COMPONENTS ANALYSIS RESULTS FOR 3-FACTOR
Component
Identified for
Item Code Deletion 1 2 3
A1_IPPA_T .665 .500 .209
A2_IPPA_T .668 .513 .112
A3_IPPA_T .673 .514 .145
A4_IPPA_T .687 .497 .174
A5_IPPA_T X .555 .114 .053
A6_IPPA_T X .596 .511 .255
A7_IPPA_T X .575 .521 .260
A8_IPPA_T .507 .685 .205
A9_IPPA_T .490 .600 .247
A10_IPPA_T .647 .568 .133
A11_IPPA_C .309 .673 .270
A12_IPPA_C X .462 .548 .156
A13_IPPA_C .264 .706 .146
A14_IPPA_C X .345 .329 .058
A15_IPPA_C .313 .678 .225
A16_IPPA_C .018 .707 .236
A17_IPPA_C .241 .680 .259
A18_IPPA_C .288 .767 .244
A19_IPPA_C .320 .700 .199
A20_IPPA_A .629 .263 .083
A21_IPPA_A .710 .288 .092
A22_IPPA_A X .336 .323 .108
A23_IPPA_A .769 .332 .102
A24_IPPA_A X .465 .572 .171
A25_IPPA_A X .480 .473 .160
A26_Inv_Aff .681 .450 .192
A27_Inv_Aff .545 .654 .191
A28_Inv_Aff .746 .342 .153
A29_Inv_Att .254 .710 .198
A30_Inv_Att .318 .749 .183
A31_Inv_Att .612 .539 .133
A32_Inv_DR .344 .642 .260
A33_Inv_DR X .538 .534 .197
A34_Inv_Dep .535 .673 .110
A35_Inv_Dep .503 .677 .197
A36_Inv_Dep .513 .667 .139
195
A37_Inv_Dep .675 .498 .154
196
A77_CLM_SS .180 .607 .366
A78_CLM_SS .656 .365 .145
A79_CLM_SS .566 .620 .207
A80_CLM_AS .648 .469 .127
A81_CLM_AS X .576 .468 .187
A82_CLM_AS .681 .475 .146
A83_CLM_AS .667 .354 .099
A84_BNS_Rel .448 .708 .215
A85_BNS_Rel X .535 .512 .106
A86_BNS_Rel .124 .609 .365
A87_YCATS_W X .357 .494 .194
A88_YCATS_W .688 .453 .174
A89_YCATS_W .334 .724 .302
A90_YCATS_W .627 .451 .280
A91_YCATS_W .689 .446 .128
A92_YCATS_W X .552 .560 .258
A93_YCATS_W .707 .376 .074
A94_YCATS_W X .461 .482 .253
A95_YCATS_W .117 .664 .339
A96_YCATS_W X .078 .429 .450
A97_YCATS_W X .503 .447 .218
A98_YCAT_A .152 .256 .733
A99_YCAT_A X .232 .140 .416
A100_YCAT_A .269 .323 .671
A101_YCAT_A .184 .182 .622
A102_YCAT_A X .299 .473 .372
A103_YCAT_A X .480 .470 .220
A104_YCAT_C .704 .042 .088
A105_YCAT_C X .395 .018 .104
A106_YCAT_C .731 .128 .138
A107_YCAT_C .728 .119 .133
A108_YCAT_C .684 .118 .135
A109_YCAT_C .624 .141 .092
A110_YCAT_C X .279 .113 .150
A111_YCAT_C .647 .046 .089
A112_YCAT_C .770 .275 .147
A113_YCAT_C X .584 .175 .152
A114_STRS .633 .564 .127
A115_STRS .348 .725 .246
A116_STRS .690 .374 .092
A117_STRS .281 .692 .277
197
A118_STRS X .055 .539 .066
A119_STRS .419 .719 .249
A120_STRS .675 .474 .139
A121_STRS .083 .643 .100
A122_STRS .780 .228 .121
A123_STRS X .320 .470 .183
A124_STRS .617 .520 .164
A125_STRS .301 .630 .307
A126_STRS .225 .769 .222
A127_STRS X -.052 .449 .319
A128_STRS X .555 .587 .132
A129_STRS .753 .344 .079
A130_NRS_A .620 .618 .187
A131_NRS_A .567 .649 .209
A132_NRS_A .609 .606 .214
A133_NRS_A .594 .622 .200
A134_NRS_A X .556 .470 .170
A135_NRS_I .313 .792 .253
A136_NRS_I .188 .682 .281
A137_NRS_I .330 .751 .222
A138_NRS_I .318 .783 .216
A139_NRS_I .295 .698 .322
A140_Sch_Ut X .381 .590 .158
A141_Sch_Ut .303 .627 .211
A142_Sch_Ut .389 .734 .200
A143_Emul .346 .678 .196
A144_Emul .477 .660 .175
A145_Emul .388 .689 .210
A146_Emul .361 .673 .154
A147_ClsSup X .588 .533 .208
A148_ClsSup .135 .775 .219
A149_ClsSup .821 .237 .108
A150_ClsSup .686 .333 .087
A151_ClsSup X .584 .499 .124
A152_ClsSup X .545 .560 .117
A153_ClsSup .628 .569 .199
A154_ClsSup .754 .417 .162
A155_ClsSup X .595 .524 .178
A156_ClsSup .670 .460 .116
A157_MePart .054 .202 .721
A158_MePart .285 .295 .643
198
A159_MePart .185 .282 .614
A160_MePart .147 .253 .751
A161_MePart .042 .337 .699
A162_MePart .301 .294 .601
A163_MePart .168 .107 .611
A164_MePart .064 .260 .651
A165_MePart .074 .305 .690
A166_MePart .172 .276 .705
A167_RAPS_E .537 .645 .216
A168_RAPS_E .778 .326 .082
A169_RAPS_E .733 .432 .147
A170_RAPS_E .543 .639 .214
199
APPENDIX K. PRINCIPAL COMPONENTS ANALYSIS RESULTS FOR 3-FACTOR
Component
Identified for
Item Code Deletion 1 2 3
A1_IPPA_T X .676 .491 .211
A2_IPPA_T X .676 .502 .121
A3_IPPA_T X .676 .510 .150
A4_IPPA_T .696 .490 .173
A8_IPPA_T X .515 .678 .205
A9_IPPA_T X .497 .594 .246
A10_IPPA_T X .657 .558 .137
A11_IPPA_C .316 .670 .274
A13_IPPA_C .272 .700 .144
A15_IPPA_C .312 .683 .222
A16_IPPA_C .025 .713 .228
A17_IPPA_C .252 .678 .255
A18_IPPA_C .289 .774 .243
A19_IPPA_C .324 .700 .194
A20_IPPA_A .628 .262 .087
A21_IPPA_A .711 .284 .105
A23_IPPA_A .773 .327 .111
A26_Inv_Aff .690 .440 .190
A27_Inv_Aff X .554 .648 .190
A28_Inv_Aff .751 .329 .161
A29_Inv_Att .261 .712 .189
A30_Inv_Att .324 .749 .174
A31_Inv_Att X .611 .533 .142
A32_Inv_DR .351 .631 .262
A34_Inv_Dep X .543 .665 .115
A35_Inv_Dep X .509 .675 .194
A36_Inv_Dep X .516 .663 .141
A37_Inv_Dep .679 .487 .163
A40_ERQ .325 .742 .160
A41_ERQ .401 .729 .180
A42_ERQ .435 .722 .198
A43_ERQ .364 .760 .203
A44_ERQ .434 .680 .195
A45_ERQ .349 .705 .208
A46_ERQ .455 .684 .184
200
A48_Friend .770 .153 .144
A49_Friend X .622 .543 .231
A50_Friend .666 .125 .128
A51_Friend .730 .154 .115
A53_Friend .609 -.006 .048
A54_Rel X .661 .573 .187
A55_Rel .406 .688 .246
A56_Rel .710 .439 .141
A57_Rel .726 .440 .161
A58_PSSM X .627 .548 .185
A60_PSSM X .606 .535 .240
A63_PSSM .474 .671 .200
A64_PSSM .732 .377 .206
A65_PSSM .672 .398 .135
A67_PSSM .778 .379 .145
A70_PSSM .734 .433 .171
A72_PSSM .680 .482 .137
A73_PSSM X .555 .639 .134
A74_PSSM X .678 .505 .193
A75_PSSM .723 .268 .125
A76_CLM_SS X .578 .628 .204
A77_CLM_SS .188 .609 .355
A78_CLM_SS .663 .355 .150
A79_CLM_SS X .576 .611 .209
A80_CLM_AS .657 .452 .133
A82_CLM_AS .688 .459 .156
A83_CLM_AS .681 .332 .108
A84_BNS_Rel .457 .704 .211
A86_BNS_Rel .130 .614 .354
A88_YCATS_W .698 .441 .177
A89_YCATS_W .339 .725 .298
A90_YCATS_W X .628 .442 .276
A91_YCATS_W .698 .428 .133
A93_YCATS_W .716 .356 .087
A95_YCATS_W .123 .668 .330
A98_YCAT_A .146 .263 .736
A100_YCAT_A .264 .326 .677
A101_YCAT_A .180 .180 .631
A104_YCAT_C .705 .035 .089
A106_YCAT_C .738 .123 .139
A107_YCAT_C .723 .123 .127
201
A108_YCAT_C .680 .121 .121
A109_YCAT_C .623 .134 .098
A111_YCAT_C .648 .043 .079
A112_YCAT_C .774 .270 .151
A114_STRS X .643 .554 .133
A115_STRS .353 .723 .245
A116_STRS .693 .371 .095
A117_STRS .285 .691 .279
A119_STRS .426 .720 .244
A120_STRS .678 .466 .145
A121_STRS .091 .639 .094
A122_STRS .787 .223 .122
A124_STRS X .620 .511 .169
A125_STRS .304 .638 .294
A126_STRS .231 .776 .213
A129_STRS .758 .336 .091
A130_NRS_A X .627 .611 .187
A131_NRS_A X .575 .641 .211
A132_NRS_A X .616 .595 .220
A133_NRS_A X .604 .615 .196
A135_NRS_I .320 .793 .245
A136_NRS_I .197 .689 .264
A137_NRS_I .330 .759 .211
A138_NRS_I .324 .786 .205
A139_NRS_I .299 .706 .311
A141_Sch_Ut .312 .619 .210
A142_Sch_Ut .392 .735 .197
A143_Emul .355 .678 .190
A144_Emul X .482 .660 .175
A145_Emul .397 .687 .209
A146_Emul .366 .676 .151
A148_ClsSup .144 .782 .207
A149_ClsSup .830 .229 .115
A150_ClsSup .688 .321 .094
A153_ClsSup X .640 .561 .194
A154_ClsSup .765 .405 .168
A156_ClsSup .676 .445 .125
A157_MePart .051 .203 .728
A158_MePart .282 .299 .647
A159_MePart .187 .280 .613
A160_MePart .143 .255 .757
202
A161_MePart .042 .340 .701
A162_MePart .295 .302 .601
A163_MePart .161 .111 .617
A164_MePart .064 .255 .665
A165_MePart .074 .301 .697
A166_MePart .168 .279 .714
A167_RAPS_E X .544 .643 .213
A168_RAPS_E .782 .320 .091
A169_RAPS_E .737 .429 .151
A170_RAPS_E X .549 .634 .218
203
APPENDIX L. PRINCIPAL COMPONENTS ANALYSIS RESULTS FOR 3-FACTOR
204
A93_YCATS_W X .351 .708 .102
A95_YCATS_W X .680 .125 .319
A98_YCAT_A .268 .138 .734
A100_YCAT_A X .330 .258 .676
A101_YCAT_A .176 .172 .639
A104_YCAT_C .044 .714 .087
A106_YCAT_C .134 .748 .132
A107_YCAT_C .134 .728 .125
A108_YCAT_C .136 .685 .111
A109_YCAT_C .144 .633 .094
A111_YCAT_C .057 .657 .070
A112_YCAT_C .275 .774 .155
A115_STRS X .721 .347 .247
A116_STRS X .380 .697 .096
A117_STRS .692 .283 .278
A119_STRS X .718 .423 .246
A120_STRS X .472 .679 .148
A121_STRS .646 .093 .084
A122_STRS .234 .791 .121
A125_STRS X .646 .303 .287
A126_STRS .785 .233 .205
A129_STRS X .341 .760 .096
A135_NRS_I X .795 .316 .245
A136_NRS_I .696 .194 .259
A137_NRS_I X .765 .330 .207
A138_NRS_I X .793 .324 .200
A139_NRS_I X .715 .300 .303
A141_Sch_Ut X .621 .311 .213
A142_Sch_Ut X .739 .390 .197
A143_Emul X .679 .353 .193
A145_Emul X .684 .395 .214
A146_Emul X .675 .365 .157
A148_ClsSup .793 .146 .197
A149_ClsSup .235 .834 .117
A150_ClsSup X .324 .686 .100
A154_ClsSup X .405 .756 .176
A156_ClsSup X .440 .666 .140
A157_MePart .201 .042 .732
A158_MePart X .302 .275 .649
A159_MePart .281 .181 .615
A160_MePart .257 .135 .760
A161_MePart X .342 .032 .703
A162_MePart X .309 .290 .599
A163_MePart .118 .162 .612
A164_MePart .253 .053 .671
A165_MePart X .301 .064 .702
A166_MePart .283 .160 .715
A168_RAPS_E X .327 .786 .097
A169_RAPS_E X .431 .735 .158
205
REFERENCES
Anastasi, A. (1982). Psychological testing (5th ed.). New York, NY: Macmillian.
Anderman, L. (1999). Classroom goal orientation, school belonging, and social goals as
Anderson, R., Manoogian, S., & Reznick, J. (1976). The undermining and enhancing of
Armsden, G. & Greenberg, M. (1978). The Inventory of Parent and Peer Attachment:
206
Baldwin, M. (1997). Relational schemas as a source of if-then self-inference procedures.
Baldwin, M. & Holmes, J. (1987). Salient private audiences and awareness of the
Baldwin, M. & Meunier, J. (1999). The cued activation of attachment relational schemas.
1130-1141.
Bao, X. & Lam, S. (2008). Who makes the choice? Rethinking the role of autonomy and
Battistich, V., Schaps, E., Watson, M., Solomon, D., & Lewis, C. (2000). Effects of the
Child Development Project on students’ drug use and other problem behaviors.
Battistich, V., Solomon, D., Kim, D., Watson, M., & Schaps, E. (1995). Schools as
207
Battistich, V., Solomon, D., Watson, M., & Schaps, E. (1997). Caring school
Baumeister, R., & Leary, M. (1995). The need to belong: Desire for interpersonal
497-529.
Belmont, M., Skinner, E., Wellborn, J., & Connell, J. (1988). Teacher as Social Context
107, 238-246.
Birch, S. & Ladd, G. (1996). Interpersonal relationships in the school environment and
Birch, S. & Ladd, G. (1997). The teacher-child relationship and children’s early school
Birch, S. & Ladd, G. (1998). Children’s interpersonal behaviors and the teacher-child
Bollen, K. (1989). Structural equations with latent variables. New York: John Wiley &
Sons.
208
Bowlby, J. (1973). Self-reliance and some conditions that promote it. In R. Gosling (Ed.),
Browne, M. & Cudeck, R. (1993). Alternative ways of assessing model fit. In K. Bollen
& J. Long (Eds.), Testing structural equation models (pp. 136–162). Newbury
Cameron, L., Ittenbach, R., McGrew, K., Harrison, P., Taylor, L., & Hwang, Y. (1997).
(Ed.), Advances in experimental social psychology (pp. 3–52). San Diego, CA:
Academic Press.
Cantor, N., Mischel, W., & Schwartz, J. C. (1982). A prototype analysis of psychological
209
Carroll, J. (1993). Human cognitive abilities: A survey of factor analytic studies.
Carvallo, M. & Gabriel, S. (2006). No man is an island: The need to belong and
32(5), 697-709.
Cattell, R. (1966). The scree test for the number of factors. Multivariate Behavioral
Research, 1, 245-276.
Cliff, N. (1983). Some cautions concerning the application of causal modeling methods,
processes across the life-span. In D. Cicchetti & M. Beeghly (Eds.), The self in
Chicago Press.
Costa, P. & McCrae, R. (1992). Normal personality assessment in clinical practice: The
Cronbach, L. (1950). Further evidence on response sets and test design. Educational and
210
Crosnoe, R., Morrison, F., Burchinal, M., Pianta, R., Keating, D., Friedman, S., Clarke-
Stewart, K., & The Eunice Kennedy Shriver National Institute of Child Health
and Human Development Early Child Car Research Network (2010). Instruction,
Davis, H. (2001). The quality and impact of relationships between elementary school
431-453.
Press.
Demanet, J. & Van Houtte, M. (2012). School belonging and school misconduct: The
differing role of teacher and peer attachment. Journal of Youth and Adolescence,
41, 499-514.
ON.pdf
211
DiStefano, C. & Motl, R. (2006). Further investigating method effects associated with
13, 440-464.
DiStefano, C. & Motl, R. (2009). Self-esteem and method effects associated with
Doumen, S., Koomen, H., Buyse, E., Wouters, S., & Verschueren, K. (2012). Teacher
motivation in education: Goals and cognitions (pp. 139-186). San Diego, CA:
Academic.
Fabrigar, L., Wegener, D., MacCallum, R., & Strahan, E. (1999). Evaluating the use of
4(3), 272-299.
212
Fairbairn, W. (1952). An object-relations theory of personality. New York: Basic Books.
Feldlaufer, H., Midgley, C., & Eccles, J. (1988). Student, teacher, and observer
perceptions of the classroom environment before and after the transition to junior
Finn, J. (1989). Withdrawing from school. Review of Educational Research, 59, 117-142.
Fiske, A. (1992). The four elementary forms of sociality: Framework for a unified theory
Clark & S. T. Fiske (Eds.), Affect and cognition: The 17th Annual Carnegie
Fiske, S. & Cox, M. (1979). Person concepts: The effect of target familiarity and
47, 136-161.
Fleming, J. (1972). Early object deprivation and transference phenomena: The working
Fraire, M., Longobardi, C., Prino, L. E., Sclavo, E., & Settanni, M. (2013). Examining
213
Freeman, T., Anderman, L., & Jensen, J. (2007). Sense of belonging in college freshmen
75(3), 203-220.
Freud, S. (1930). Civilization and its Discontents. (J. Riviere, Trans.) London: Hogarth
Press.
Fromm, E. (1956). The Art of Loving. New York, NY: Harper & Row.
Bulletin, 2, 66-70.
Goodenow, C. (1992). Strengthening the links between educational psychology and the
Goodenow, C. & Grady, K. (1993). The relationship of school belongingness and friends’
214
Gregoriadis, A. & Tsigilis, N. (2008). Applicability of the Student-Teacher Relationship
Grusec, J. & Goodenow, J. (1994). Impact of parental discipline methods on the child’s
Gutman, L. & Midgley, C. (2000). The role of protective factors in supporting the
Hamre, B. & Pianta, R. (2001). Early teacher-child relationships and the trajectory of
children’s school outcomes through eighth grade. Child Development, 72, 625-
638.
Hamre, B. & Pianta, R. (2005). Can instructional and emotional support in the first-grade
Henricsson, L. & Rydell, A. (2004). Elementary school children with behavior problems:
Higgins, E. & Bargh, J. (1987). Social cognition and social perception. Annual Review
215
Higgins, E., Rholes, W., & Jones, C. (1977). Category accessibility and impression
Hodgins, H., Koestner, R., & Duncan, N. (1996). On the compatibility of autonomy and
Horan, P, DiStefano, C., & Motl, R. (2003). Wording effects in self-esteem scales:
455.
Horn, J. & Cattell, R. (1966). Refinement and test of the theory of fluid and crystalized
Horney, K. (1945). Our Inner Conflicts: A constructive Theory of Neurosis. New York:
Norton.
House, J., Landis, K., & Umberson, D. (1988). Social relationships and health. Science,
241, 540-545.
9(2), 191-204.
Hu, L. & Bentler, P. (1999). Cutoff criteria for fit indexes in covariance structure
Modeling, 6, 1–55.
216
IRRE (1998). Research Assessment Package For Schools (RAPS): Student Supports and
Johnson, D. & Johnson, R. (1983). Social interdependence and perceived academic and
personal support in the classroom. The Journal of Social Psychology, 120, 77-82.
Johnson, D., Johnson, R., Buckman, L., & Richards, P. (1985). The effect of prolonged
Johnson, L., Ketring, S., & Abshire, C. (2003). The revised inventory of parent
25(3), 333-349.
Joreskog, K. (1999). How large can a standardized coefficient be? Retrieved September
izedCoefficientbe.pdf.
Kaiser, H. (1958). The varimax criterion for analytic rotation in factor analysis.
Kennedy, G. & Tuckman, B. (2013). An exploration into the influence of academic and
217
Kester, V. (1994). Factors that affect African-American students’ bonding to middle
Klem, A. & Connell, J. (2004). Relationships matter: Linking teacher support to student
Kohl, D., Recchia, S., & Steffgen, G. (2013). Measuring school climate: An overview of
Koomen, H., Verschueren, K., van Schooten, E., Jak, S., & Pianta, R. (2012). Validating
invariance across child gender and age in a Dutch sample. Journal of School
Ladd, G., Birch, S., & Buhs, E. (1999). Children’s social and scholastic lives in
1400.
Ladd, G. & Burgess, K. (2001). Do relational risks and protective factors moderate the
Laguardia, J., Ryan, R., Couchman, C. & Deci, E. (2000). Within-person variation in
79, 367-384.
218
Lisonbee, J., Mize, J., Payne, A., Granger, D. (2008). Children's cortisol and the quality
teachers. In R. C. Pianta (Ed.), Beyond the parent: The role of other adults in
children’s lives (Vol. 57, pp. 81-107). San Francisco, CA: Jossey-Bass.
Lynch, M. & Cicchetti, D. (1997). Children’s relationships with adults and peers: An
Maassen, G. & Bakker, A. (2001). Suppressor variables in path models: Definitions and
their teachers: Associations with child and classroom context variables. Journal of
219
Markus, H. (1977). Self-schemata and processing information about the self. Journal of
819.
Marsh, H., Hau, K., & Wen, Z. (2004). In search of golden rules: Comment on hypothesis
testing approaches to setting cut-off values for fit indexes and dangers in
11, 320-341.
Matzye, C. (1995). Parental involvement in middle school. ERIC Document ED 385 365.
Maulana, R., Opdenakker, M., Stroet, K., & Bosker, R. (2013). Changes in teachers’
involvement versus rejection and links with academic motivation during the first
McCallum, R. & Bracken, B. (1993). Interpersonal relations between school children and
McCrae, R. & Costa, P. (1987). Validation of the five-factor model of personality across
instruments and observers. Journal of Personality and Social Psychology, 52, 81-
90.
220
Midgley, C., Feldlaufer, H., & Eccles, J. (1989). Student/teacher relations and attitudes
toward mathematics before and after the transition to junior high school. Child
Milatz, A., Gluer, M., Harwardt-Heinecke, E., Kappler, G., & Ahnert, L. (2014). The
Miller, G. (1956). The magical number seven, plus or minus two: Some limits on our
University Press.
Mitchell-Copeland, J., Denham, S., & DeMulder, E. (1997). Q-Sort assessment of child-
221
Oldfather, P. & Dahl, K. (1994). Toward a social constructivist reconceptualization of
intrinsic motivation for literacy learning. Journal of Reading Behavior, 26, 139-
158.
Pace, C., Martini, P., & Zavattini, G. (2011). The factor structure of the Inventory of
Palermo, F., Hanish, L., Martin, C., Fabes, R., & Reiser, M. (2007). Preschoolers’
academic readiness: What role does the teacher-child relationship play? Early
Patrick, H., Kaplan, A., & Ryan, A. (2011). Positive classroom motivational
Patrick, H., Ryan, A., & Kaplan, A. (2007). Early adolescents’ perceptions of the
222
Pianta, R. (2006). Teacher-child relationships and early literacy. In D. Dickinson & S.
Neuman (Eds.), Handbook of early literacy research (Vol. 2, pp. 149-162). New
Pianta, R., Belsky, J., Vandergrift, N., Houts, R., & Morrison, F., (2008). Classroom
adjusting to school. In R. C. Pianta (Ed.), Beyond the parent: The role of other
Pietromonoco, P. & Barrett, L. (2000). The internal working models concept: What do
McGhee, & D. D. Clarke (Eds.), Accounting for relationships (pp. 175–191). New
York: Methuen.
Podsakoff, P., MacKenzie, S., Lee, J., & Podsakoff, N. (2003). Common method biases
223
Preston, C. & Colman, A. (2000). Optimal number of response categories in rating scales:
Reeve, J. & Jang, H. (2006). What teachers say and do to support students' autonomy
Resnick, M., Bearman, P., Blum, R., Bauman, K., Harris, K., Jones, J., Tabor, J.,
Beuhring, T., Sieving, R., Shew, M., Ireland, M., Bearinger, L., & Udry, J.
(1997). Protecting adolescents from harm: Findings from the national longitudinal
278(10), 823-832.
Roeser, R., Midgley, C., & Urdan, T. (1996). Perceptions of school psychological
224
Rudasill, K. (2011). Child temperament, teacher-child interactions, and teacher-child
Russell, D. (2002). In search of underlying dimensions: The use (and abuse) of factor
Ryan, R. (1993). Agency and organization: Intrinsic motivation, autonomy, and the self
Ryan, R. (1995). Psychological needs and the facilitation of the integrative processes.
Ryan, R., Connell, J., & Deci, E. (1985). A motivational analysis of self-determination
Ryan, R. & Deci, E. (2000). Self-determination theory and the facilitation of intrinsic
68-78.
Ryan, R., & Grolnick, W. (1986). Origins and pawns in the classroom: Self-report and
225
Ryan, R. & Powelson, C. (1991). Autonomy and relatedness as fundamental to
Ryan, R., Stiller, J., & Lynch, J. (1994). Representations of relationships to teachers,
Ryzin, M., Gravely, A., & Roseth, C. (2009). Autonomy, belongingness, and engagement
Sandler, J. & Rosenblatt, B. (1962). The concept of the representational world. The
Research, 8, 3-15.
Press.
Schank, R. & Abelson, R. (1977). Scripts, plans, goals, and understanding. Hillsdale,
NJ: Erlbaum.
Schlomer, G., Bauman, S., & Card, N. (2010). Best practices for missing data
57(1), 1-10.
294-309.
226
Shochet, I. & Smith, C. (2014). A prospective study investigating the links among
Shochet, I., Smith, C., Furlong, M., & Homel, R. (2011). A prospective study
595.
teacher behavior and students engagement across the school year. Journal of
Skinner, E., Furrer, C., Marchand, G., & Kinderman, T. (2008). Engagement and
Skues, J., Cunningham, E., & Pokharel, T. (2005). The influence of bullying behaviours
Solomon, D., Battistich, V., Kim, D., & Watson, M. (1997). Teacher practices associated
Education, 1, 235-267.
Solomon, D., Battistich, V., Watson, M., Schaps, E., Lewis, C. (2000). A six-district
study of educational change: Direct and mediated effects of the child development
227
Solomon, D., Watson, M., Battistich, V., Schaps, E., & Delucchi, K. (1996). Creating
Solomon, D., Watson, M., Delucchi, K., Schaps, E., & Battistich, V. (1988). Enhancing
Srull, T. & Wyer, R. (1979). The role of category accessibility in the interpretation of
Stipek, D. & Miles, S. (2008). Effects of aggression on achievement: Does conflict with
Valiente, C., Lemery-Chalfant, K., Swanson, J., & Reiser, M. (2008). Prediction of
Van Ryzin, M., Gravely, A., & Roseth, C. (2009). Autonomy, belongingness, and
228
Webb, M. & Neuharth-Pritchett, S. (2011). Examining factorial validity and
173-182.
Wentzel, K. (1998). Social relationships and motivation in middle school: The role of
Wentzel, K., Baker, S., & Russell, S. (2012). Young adolescents’ perceptions of teachers’
and peers’ goals as predictors of social and academic goal pursuit. Applied
Winnicott, D. (1965). The maturational process and the facilitating environment. New
229
Ye, F. & Wallace, T. (2014). Psychological Sense of School Membership scale: Method
You, S., Ritchey, K., Furlong, M., Shochet, I., & Boman, P. (2011). Examination of the
Zumbrunn, S., McKim, C., Buhs, E., & Hawley, L. (2014). Support, belonging,
230