Download as pdf or txt
Download as pdf or txt
You are on page 1of 14

Studies in Educational Evaluation 38 (2012) 107–120

Contents lists available at SciVerse ScienceDirect

Studies in Educational Evaluation


journal homepage: www.elsevier.com/stueduc

Learning environment and students’ mathematics attitude


Machteld Vandecandelaere *, Sara Speybroeck, Gudrun Vanlaar, Bieke De Fraine, Jan Van Damme
K.U.Leuven (University of Leuven), The Education and Training Research Group, Centre for Educational Effectiveness and Evaluation, Dekenstraat 2/3773, 3000 Leuven, Belgium

A R T I C L E I N F O A B S T R A C T

Article history: This study investigated the association between students’ perception of the learning environment and
Received 24 February 2012 three aspects of their mathematics attitude: ‘mathematics academic self-concept’, ‘enjoyment of
Received in revised form 22 August 2012 mathematics’ and ‘perceived value of mathematics’. The focus was on the association of students’
Accepted 2 September 2012
mathematics attitude with four dimensions in the learning environment: the extent to which the teacher
‘motivates to exert learning effort’, ‘activates towards self-regulated learning’, ‘gives feedback and
Keywords: coaches’, and ‘structures and steers’. Data were obtained from an extended version of the international
Mathematics attitude
Trends in Mathematics and Science Study (TIMSS) of 2003. Multilevel analysis on a sample of 4354
Learning environment
Multilevel analysis
eighth grade students in 228 classes in 119 schools in Flanders (Belgium) indicated that the learning
Trends in Mathematics and Science Study environment plays a significant role in the enjoyment of mathematics. This while the mathematics
2003 academic self-concept and the perceived value of mathematics are insensitive for aspects in the learning
environment.
ß 2012 Elsevier Ltd. All rights reserved.

Introduction Mathematics attitude

Together with the rise of a positive psychology, the importance Attitude has been a difficult concept to define, primarily because
of a positive attitude towards educational matters has gained more it has been defined by so many, but also because it is considered as a
attention in recent years (e.g., Marsh & Craven, 2006; Seligman & multidimensional concept (Di Martino & Zan, 2010; Hanulla, 2002;
Csikszentmihalyi, 2000). Research suggested repeatedly that Kadijevich, 2006; Meelissen & Doornekamp, 2004; Simonson &
mathematics attitude is a critical construct related to learning Maushak, 2001; Tapia & Marsh, 2005; Watt, 2000). A frequently
(e.g., Singh, Granville, & Dika, 2002). The extent to which students cited definition is that of Zimbardo and Leippe (1991) who define
find mathematics enjoyable, place value on mathematics and attitude as ‘‘an evaluative disposition towards some object based
believe that it is important for success in school and future upon cognitions, affective reactions, behavioural intentions, and
aspirations, affects the students’ motivation to learn (Ismail, 2009; past behaviour that can influence cognitions, affective responses,
Mettas, Karmiotis, & Christoforou, 2006; Middleton & Spanias, and future intentions and behaviours’’ (p. 32). Overall, different
1999). Other studies have identified attitudinal and affective categorisations of attitude are tantamount to three interwoven
variables, such as self-concept, confidence, mathematics interest dimensions: affective responses, cognitions and behavioural inten-
and motivation as salient predictors of achievement in mathemat- tions (Ajzen, 1988; Di Martino & Zan, 2010; Gómez-Chacón, 2000;
ics (Reynolds & Walberg, 1992; Schreiber, 2000; Singh et al., 2002; Ruffell, Mason, & Allen, 1998). The affective dimension involves the
Thorndike-Christ, 1991). Moreover, these variables are associated student’s evaluation of a situation, object or person, the extent to
with mathematics avoidance, which affects long-term achieve- which the student likes it, and the student’s emotional response to it.
ment and career aspirations in the mathematics domain (Crombie The cognitive dimension consists of a student’s beliefs about the
et al., 2005; Helmke, 1989; Reynolds & Walberg, 1992). In short, it situation, object or person. Finally, the behavioural intention
is clear that the attitude towards mathematics is a vital matter in dimension contains the plans to act in a certain way. However,
mathematics education. Therefore, this study seeks to identify the the behavioural dimension is not always seen as a dimension of
role of the learning environment in enhancing the attitude towards attitude as such (Daskalogianni & Simpson, 2000). Studies on
mathematics. mathematics attitude acknowledge the multifaceted structure of
attitude. For example, Watt (2000) argued that mathematics
attitude consists of perceived talent, expected success, effort
* Corresponding author. Tel.: +32 16 32 57 80; fax: +32 16 32 58 59. requirement, difficulty, interest and utility. Hanulla (2002) distin-
E-mail addresses: Machteld.Vandecandelaere@ppw.kuleuven.be
guished four evaluative processes: emotions aroused in the
(M. Vandecandelaere), Sara.Speybroeck@ppw.kuleuven.be (S. Speybroeck),
Gudrun.Vanlaar@ppw.kuleuven.be (G. Vanlaar), Bieke.DeFraine@ppw.kuleuven.be situation, emotions associated with the stimuli, expected conse-
(B. De Fraine), Jan.VanDamme@ppw.kuleuven.be (J. Van Damme). quences, and the act of relating the situation to personal values.

0191-491X/$ – see front matter ß 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.stueduc.2012.09.001
108 M. Vandecandelaere et al. / Studies in Educational Evaluation 38 (2012) 107–120

Further, Tapia and Marsh (2005) found self-confidence, value of students with lowly educated parents (Kogce, Yildiz, Aydin, &
mathematics, enjoyment of mathematics and motivation to be the Altindag, 2009; Van Damme et al., 2004).
main components of the mathematics attitude. Meelissen and With regard to the association with (2) the outcomes, we also
Doornekamp (2004) reported four components, i.e., self-confidence, accounted for intake characteristics of which an association with
enjoyment, value and support. More recently, Di Martino and Zan mathematics attitude can be expected, based on previous research.
(2010, 2011) presented a three-dimensional model of mathematics These are: gender, language, country of origin, study section and
attitude, consisting of the student’s vision of mathematics, perceived mean cognitive abilities of the class group. Concerning the relation
competence and an emotional dimension. Their model explicitly with gender, several studies have demonstrated a slightly more
takes into account beliefs (about self and mathematics) and negative attitude towards mathematics for girls (Chouinard,
emotions, as well as the interplay between them. Vezau, Bouffard, & Jenkins, 1999; Frost, Hyde, & Fennema, 1994;
In the present study, mathematics attitude was measured using Leder, 1995; Meelissen & Doornekamp, 2004; Van Damme et al.,
the student questionnaire of the international Trends in Mathe- 2004). However, according to Chouinard et al. (1999) the gender
matics and Science Study (TIMSS) of 2003. TIMSS items reflect a gap tends to diminish or disappear from the age of 14. Meelissen
consensus among the representatives of the participating coun- and Doornekamp (2004) found a more positive attitude for eight
tries. As a consequence, the items adopted in the TIMSS grade boys concerning their self-confidence and the perceived
questionnaire are not explicitly based on a theoretical framework. value of mathematics, yet no significant difference concerning the
However, Kadijevich (2008) explains how three dimensions can be enjoyment of mathematics. Kogce et al. (2009) found no significant
distinguished in the items on mathematics attitude: ‘‘(1) self- gender effect on the mathematics attitude in grades 6, 7 and 8.
confidence denotes perceived ease, or difficulty, of learning With regard to the home language of students, Tymms (2001)
mathematics; (2) liking mathematics stands for student’s affective, found that children whose first language was not English are
emotional and behavioral reactions concerning liking, or disliking, marginally more positive towards mathematics. Likewise, Van
mathematics; (3) usefulness of mathematics denotes student’s Damme et al. (2006) found that students with Dutch as the
beliefs concerning the contribution of mathematics to his/her exclusive home language have a lower self-concept than foreign
educational and vocational performance’’ (p. 330). In the method language speakers. Further, research regarding the effect of the
section, the operationalisation of the mathematics attitude is country of origin of the student has produced mixed results (e.g.,
described in more detail. Garaway, 1994; Ma & Kishore, 1997; Randhawa & Gupta, 2000)
with some studies finding significant attitudinal differences
Student characteristics and mathematics attitude between national groups while others do not. We also accounted
Students are not randomly assigned to schools and classes. The for the study section (subject choice for a limited amount of hours
assignment depends on nonrandom processes, including socioeco- each week) as Van Damme et al. (2004) reported that students in
nomic and racial residential segregation (Massey & Denton, 1988). the general and technical study section have a more negative
As a consequence, schools and classes vary on the intake mathematics attitude compared to students in the classical study
characteristics of their students (Raudenbush & Willms, 1995). section. Further, it has been shown that not only the cognitive
Ignoring this fact results in biased estimates, i.e., found differences in abilities of the student but also the average cognitive abilities of
the mathematics attitude would be ascribable to intake character- the group affect students’ academic self-concept. This is known as
istics, rather than to the impact of the learning environment. To the Big-fish-little-pond effect (Marsh, Trautwein, Lüdtke, Baumert,
account for non-randomness, research should include students’ & Köller, 2007). Marsh et al. (2007) point out that students with the
background to such a degree that school or class enrolment is same cognitive abilities have a higher academic self-concept when
strongly ignorable. This implies considering covariates related to they attend a school where the average cognitive abilities are low.
both (1) school and class membership and (2) the outcome (Kennedy
& Mandeville, 2000). With regard to (1) school and class member- Changeability of mathematics attitude
ship, research has shown that general measures of verbal or In this study, we were interested in the degree to which
quantitative ability and measures of the family background capture malleable aspects in the learning environment are associated with
most of the differences between schools and classes regarding their mathematics attitude. This involves the assumption that the
intake (Kennedy & Mandeville, 2000; Willms, 2002). Therefore, we attitude towards mathematics can be influenced and accordingly
accounted in the present study for the cognitive abilities of the calls for a clarification on the formation and modification of
students and two indicators of the social economic family attitude. Attitude is formed both directly and indirectly through
background, which are parental education level and home posses- situational stimuli or events in the environment (Simonson &
sions.1 These two variables were used in the international TIMSS Maushak, 2001; Zimbardo & Leippe, 1991). Unlike personality,
data collection as indicators of the home environment and the experiences may affect attitude (Tesser, 1993). Several attempts
socioeconomic status (SES) (Mullis, Martin, & Stemler, 2000). These have been made to point out the possibility of enhancing and
variables are not only related with the school and class enrolment modifying affective and attitudinal factors by curricular and
but also with the outcome, i.e., mathematics attitude. Previous instructional classroom practices. Papanastasiou (2008) demon-
studies have revealed that students with higher cognitive abilities strated a positive association between a clear, well organised
tend to have a more positive attitude towards mathematics than teaching approach that keeps students involved and student’s
students with lower cognitive abilities (e.g., Ma, 1997; Van Damme, mathematics attitude. Based on an extensive literature review,
Opdenakker, & Van den Broeck, 2004). Pahlavan, Farzad, and Naderi Simonson and Maushak (2001) argue that authentic situations
(2007) found a positive correlation between the socioeconomic which are relevant to the students, as well as the experience of
status and the mathematics attitude in grade 8. Furthermore, several purposeful emotional involvement are likely to change attitude in
studies have revealed that students with highly educated parents the direction advocated in the situation. Students are motivated
are more likely to have a positive attitude towards mathematics than and react positively when the learning environment includes the
discovery of useful new information about a topic. Moreover, post-
1
instruction activities, such as follow-up discussions, were found to
In TIMSS 2003 there were 16 ‘‘possessions’’ questions relating to the following
items: calculator, computer, study desk/table, dictionary, electricity, running tap
be a powerful way to produce attitude changes (Simonson &
water, television, video player, CD player, radio, own bedroom, water flushed Maushak, 2001). Correspondingly, Tobias and Weissbrod (1980)
toilets, motor car, own bicycle, telephone, fridge. investigated how a curriculum that focuses on conceptualising and
M. Vandecandelaere et al. / Studies in Educational Evaluation 38 (2012) 107–120 109

creating relevance and meaning, improves students’ motivation to Activate towards self-regulated learning
learn mathematics. Further, Di Martino and Zan (2010) stress the This dimension incorporates all activities ‘‘which aim at engaging
importance of intervening on a negative mathematics attitude by pupils in learning tasks in which they have to link/question their
activities which are focussed on mathematical processes (e.g., prior knowledge with new information to be able to construct
students’ vision of mathematics and perceived competence) rather meaning’’ (Schelfhout et al., 2006, p. 478). The learning assignments
than on the mathematical products itself. Finally, Van Damme et al. should be sufficiently complex to activate students towards self-
(2004) demonstrated that differences in mathematics attitude can regulated learning, though not too complex to avoid misconceptions
be explained by the way in which students experience the teaching and demotivation. These tasks should also be supported by
of mathematics. activating and structuring feedback, which constitutes the third
The research to date has tended to focus on the association dimension in the model. Another important feature in this
between the learning environment with either an undifferentiat- dimension is to create co-operative learning environments in order
ed operationalisation of attitude or a single subdimension of to activate students to articulate their thoughts, connecting them to
attitude. The present study seeks to extent the existing body of what has been discussed in the group before (Schelfhout et al., 2006).
knowledge by investigating the association between the four Studies on the effect of activation towards self-regulated learning
dimensions in the ‘Equilibrium model for creating a powerful have shown the importance of activating teaching activities to reach
learning environment’ (Schelfhout, 2003; Schelfhout, Dochy, in-depth learning and to construct a well-organised body of
Janssens, Struyven, & Gielen, 2006) and ‘mathematics academic knowledge in a more active way (e.g., Biemans, 1997; Bonwell &
self-concept’, ‘enjoyment of mathematics’ as well as ‘perceived Eison, 1992; Schunk, 2001). Moreover, a study by Schelfhout, Van
value of mathematics’. Landeghem, Van den Broeck, and Van Damme (2007) revealed that
activating students towards self-regulated learning is positively
Equilibrium model for creating powerful learning environments related to mathematics achievement.

Based on an extensive literature study, Schelfhout (2003) and Give feedback and coach
Schelfhout et al. (2006) present a model with four dimensions for In this third dimension it is stressed that feedback should be
creating powerful learning environments. The dimensions were given before, during and after learning tasks, as well as after
operationalised into concrete educational activities which are assessments (Schelfhout et al., 2006). It is important that students
considered to be indicators of effective teaching. Teachers need to are aware of the necessary knowledge and skills to be able to
find an effective and manageable balance between these teaching handle learning tasks that can lead to in-depth learning (Prawat,
activities, taking into account the available resources, the qualities 1989). During learning tasks, the learner needs help and limited
of both the students and the teachers and the learning goals. hints to take the next self-regulated step in his or her learning
Therefore, the authors call it the ‘Equilibrium model’. process (Corno & Randi, 1997; Vygotsky, 1978). After a task or
In the present study, the four dimensions of the learning assessment, the learners should be able to check if the given
environment were measured through the students’ perceptions, solution is correct and they should be activated to use the feedback
which were aggregated to a class level. There is indeed some to reflect on the causes of mistakes (Good & Brophy, 2000; Winne,
evidence that student perceptions are more important than 1997).
objective measurements of the environment (e.g., Bennet & Ward,
1993; McDowell, 1995; Van Damme et al., 2004). A study of De Structure and steer
Jong and Westerhof (2001) demonstrated that aggregated student The authors distinguish a fourth dimension ‘‘to stress the
ratings provide a good assessment of teacher behaviour. Moreover, importance of all these educational activities which are needed to
the authors found a higher predictive validity of aggregated deliberately organise the former three main activities over a longer
student ratings compared to external observations when subject period of teaching with the goal of deep level learning and transfer’’
specific motivation is taken as a dependent variable. In the (Schelfhout et al., 2006, p. 480). Students expect from the teaching
following paragraphs, the four interconnected dimensions of the approach that they will be able to understand the subject matter
‘Equilibrium model’ are described briefly. and to solve the learning assignments (Good & Brophy, 2000). The
teacher should work out an appropriate sequence of learning tasks
Motivate to exert learning and coaching activities to reach in-depth learning (Reigeluth,
This first group of educational activities provides a focal point 1999). This can be realised more easily in traditional education
while creating powerful learning environments. The ‘‘teaching settings but should also be met when using more complex learning
activities should foster the motivation of students to exert effort to tasks which activate towards self-regulated learning.
engage in learning activities and to sustain these efforts’’
(Schelfhout et al., 2006, p. 476). First, it is important to motivate Purpose of the present study
by arousing interest. The authors claim that the teacher should use
the students’ interests as starting points, for instance by giving As students’ attitude towards mathematics is an important
examples from the actuality or the environment of the students topic to be considered in mathematics education, the question
and by capitalising on what people want to learn. Second, it is arises as to how teachers can act to stimulate a more positive
important to create and maintain motivation for learning by attitude towards mathematics in their students. Based on the
creating the will to acquire new knowledge and skills and to keep ‘Equilibrium model for creating powerful learning environments’,
these learning efforts on the same high level. Research has shown we hypothesised that there is an association between the extent to
the importance of motivating to exert learning in order to provoke which the teacher ‘motivates to exert learning’, ‘activates towards
deep level learning (e.g., Boekaerts, 2001; Stipek, 2002; Utman, self-regulated learning’, ‘gives feedback and coaches’ and ‘struc-
1997). Creating intrinsic motivation for in-depth learning (Stipek, tures and steers’ and three components of the mathematics
2002) and extrinsic motivation that can be internalised (Ryan & attitude: ‘mathematics academic self-concept’, ‘enjoyment of
Deci, 2000) are important features within this dimension. The mathematics’ and ‘perceived value of mathematics’. Because we
teacher can help students to keep learning efforts on a same high were interested in the net-effect of the learning environment, we
level by providing a pleasant variety of educational approaches accounted for variables that are beyond the control of the teacher:
(Schelfhout et al., 2006). student characteristics and the average cognitive abilities of the
110 M. Vandecandelaere et al. / Studies in Educational Evaluation 38 (2012) 107–120

class group. Therefore, the current study addresses the relation the subject matter and beliefs about the own subject-related
between the learning environment and mathematics attitudes competence. The enjoyment of mathematics reflects the affective
after controlling for these covariates. dimension of attitude. The three dimensions resemble these in the
international TIMSS questions on mathematics attitude (Kadije-
Method vich, 2008) and correspond to the characterisation of mathematics
attitude by Di Martino and Zan (2010, 2011).2 Summarising,
Data although the three attitude components in our study are initially
based on the available TIMSS data in Flanders, they echo a more
The data in this study were drawn from the Flemish data in general framework as found in the research literature on the
TIMSS 2003 in grade 8. This dataset includes information on attitude towards mathematics. An overview of the questionnaire
students’ mathematics attitude, as well as on their background, items for each component is shown in Appendix 1. Twelve of these
teacher and classroom. The data were collected by means of items were part of the international measurement in TIMSS 2003
questionnaires, administered to students and teachers. In addition and were answered on a four-point Likert scale (1 = Agree a lot to
to the international data collection, some supplementary data 4 = Disagree a lot). Fifteen items were part of the Flemish extension
were gathered in Flanders. The data collection was extended with a and were answered on a five-point Likert scale (1 = Agree a lot to
cognitive abilities test, additional questions on the mathematics 5 = Disagree a lot).3 The responses on positively formulated items
attitude, a question about the study section and questions about were recoded, so that high scores indicated higher levels of a
the way in which students perceive their learning environment. positive mathematics attitude. The internal consistency coeffi-
Furthermore, a questionnaire for the parents was added, with cients (Cronbach’s a) of the scales ‘mathematics academic self-
questions regarding their education level. concept’, ‘enjoyment of mathematics’ and ‘perceived value of
mathematics’ were .92, .93 and .83 respectively. The three
Sample components formed the three dependent variables in our study.
Descriptive statistics are shown in Appendix 3.
In Grade 8 in Flanders, students have to choose between the
general track, which attracts more than 85% of the students, and Student characteristics
the vocational track. In this study, only students in the general We accounted for the following student level variables:
track were considered. Whereas most countries participating in cognitive abilities, parental education level, home possessions,
TIMSS 2003 selected only one class in each school, Flanders gender, language, country of origin and study section (see
selected two classes in each school. The data consequently had a Appendix 3).
three-level structure, i.e., students nested within classes within The cognitive abilities were measured by means of the ‘Non-
schools. The resulting sample consisted of 4354 students, nested verbal test battery for the seventh and the eighth grade’ (Demeyer,
within 228 classes, nested within 119 schools. For a detailed 1999). This test consists of two numerical subtests (numerical
description of the sample procedure in TIMSS 2003, see Foy and symbols and numerical approximations) and two spatial subtests
Joncas (2004). (folded squares and cross-sections), each containing 20 items
(Demeyer, 1999).
Variables Both the parental education level and the home possessions
were considered as indicators of the students’ socioeconomic
Mathematics attitude status. The parental education level was retrieved from the parent
As addressed in the introduction, the operationalisation of the questionnaire. The index was initially based on the education level
mathematics attitude was derived from the data from the student of the mother. When this information was missing, the education
questionnaire in TIMSS 2003. Because we were interested in a level of the father was taken into account.4 The education level of
more comprehensive measure of the mathematics attitude, in
Flanders (i.e., the Dutch speaking part of Belgium), the student 2
More details about the psychometric qualities of the instrument are given in the
questionnaire was extended with items of the Dutch ‘Attitude method section.
3
Scale towards Mathematics’ (Martinot, Kuhlemeier, & Feenstra, TIMSS 2003 made extensive use of a 4-point Likert scale as the small number of
response categories improves the ease of administration. In the Flemish extension, a
1988). This instrument consists of four scales: ‘pleasure’, ‘confi- five-point Likert scale was used for several reasons. Previous research has shown
dence’, ‘relevance’ and ‘effort/interest’. Except for the latter scale, that a higher number of response options increases the reliability (Cox, 1980).
the scales reflect the dimensions as mentioned by Kadijevich Fewer scale points lose information and tend to produce more cognitive overload
(2008). As a consequence of both the limited theoretical (Groves et al., 2011) Moreover, whereas four categories force the respondent to
agree or disagree, an odd number of response categories allows the respondent to
framework of the international TIMSS items and the Flemish
have a neutral attitude (Allen & Seaman, 2007). However, research suggested that
extension resulting in a new combined questionnaire, we were the overall difference in responses between a 4-point and a 5-point Likert scale is
challenged to use data that can be considered reliable and valid. negligible (Armstrong, 1987). Principal component analysis makes no assumptions
Following Tahar, Ismail, Zamani, and Adnan (2010) a principal regarding the scale length, therefore, the combination of 4-point and 5-point
component analysis was carried out to determine the important questions was legitimate. The calculation of the component score is based on the
loadings of each item for that particular component (Stern, 2009).
components of students’ attitudes towards mathematics, as 4
As Jones (2001) argues, there is no single measure of the parental education
measured by the Flemish and international items. This resulted level that is generally accepted as the most appropriate measure. However, the
in a three-dimensional structure of attitude. The mathematics mother’s education level has generally been used (e.g., Reynders, Nicaise, & Van
academic self-concept concerns the student’s perceptions of his Damme, 2005), based on the argument that ‘‘that mothers (compared to fathers) are
more involved with the socialisation of a child; they spend more time with the child,
capability to master the subject matter and to do well in
spend more time reading and helping with the child’s homework, and generally are
mathematics. The enjoyment of mathematics contains the extent more aware of the child’s world at school’’ (Marks, McMillan, Jones, & Ainley, 2000,
to which the student enjoys mathematics classes and the subject p. 15). Besides, most often the effects of mother’s and father’s education appear to
matter itself. Finally, the perceived value of mathematics refers to be comparable (Marks et al., 2000). Concerning missing data, Marks et al. (2000)
the beliefs the student holds about the importance of mathematics argue that there is no advantage in using one or the other because both attract
similar levels of missing data. They point out that mother’s education has generally
in every day and later life. The mathematics academic self-concept been used in preference to father’s education, for the reasons outlined above. In
and the perceived value of mathematics reflect the cognitive cases where mother’s education was missing, father’s education was used. The
dimension of attitude, more specifically, they refer to beliefs about same approach was used in our study.
M. Vandecandelaere et al. / Studies in Educational Evaluation 38 (2012) 107–120 111

the parents was high when the mother (or father when missing) dimension of the learning environment, these factor scores were
finished higher education successfully (coded as 3). The education aggregated within classes. The four aggregated variables correlated
level was moderate when the mother (or father when missing) highly (r > .90), indicating that classes who scored high for one
finished only secondary education successfully (coded as 2). The dimension of teaching activities also tended to score high for the
education level was low when the mother (or father when missing) other dimensions. Indeed, the teaching activities in the four
had no secondary education diploma (coded as 1). dimensions are all considered to be indicators of effective teaching.
The home possessions were measured by means of 16 questions Because of the high intercorrelations, the effects of the four
about possessions at home (e.g., a computer, a spare room, etc.).2 dimensions were analysed separately for each component of the
Items on which the student answered ‘yes’, were coded as 1; ‘no’ mathematics attitude to avoid multicollinearity problems (O’Brien,
were coded as 0. The yes-answers were summed and the 2007).
unweighted sum was divided by the number of items answered
by the student. This quotient was multiplied by 100. Data analyses
Gender was included as a dichotomous variable.
The students’ home language was also included in our study as a For each of the three components of the mathematics attitude a
dichotomous variable: students who mostly or always speak Dutch series of multilevel analyses was carried out, taking into account
and students who rarely or never speak Dutch at home. the hierarchical structure of the data. A three-level unconditional
The student questionnaire also elicited information about the model, or null model, was specified for each component to address
country of origin of both the students and their parents. This the variability ascribed to each level of nesting. Hereby the
variable was included as a categorical variable. Students were variance was divided in three components, situated on the student,
considered Belgian when both themselves and their parents were class and school level. Subsequently, the intra class correlation
born in Belgium. Students who were not born in Belgium were coefficients (ICCs) were calculated (Hox, 2010).
considered as first generation immigrants. Students who were To ease the interpretation, all continuous predictors were
born in Belgium, but one or two of their parents were not, were centred around their grand mean (Hox, 2010). Next, for each
considered as second generation immigrants. component of the mathematics attitude, six models were
In Grade 8, students have to choose for a study section for a specified. In Model 1, the student level predictors were included.
limited amount of hours per week. The options are: ‘classical Subsequently, the mean cognitive abilities of the class were
languages with Greek and Latin’, ‘general subjects’ and more included in Model 2. Next, the variables ‘motivate’, ‘activate’,
‘technical options’. This variable was included as a categorical ‘feedback’ and ‘structure’ were each separately added to Model 3
variable. to 6 respectively. As there were no research questions regarding
between class or school differences in the predictive power of the
Class characteristics independent variables, the random parts of the slopes were not
At the class level, we incorporated the mean cognitive abilities estimated, and thus all models reported are random-intercept
of the class group and the averaged perceptions within each class only models.
group on four scales concerning the learning environment. The estimated fixed parameters were tested by a two-sided
The mean cognitive abilities of the class group were calculated by Wald test on the 5% significance level (Snijders & Bosker, 1999).
taking the average of the cognitive abilities test scores of the Model improvement was assessed by studying the decrease in
students in that class. deviance values of the different models. The decrease in deviance
Based on the ‘Equilibrium model for creating powerful learning was tested with a Likelihood ratio test (Hox, 2010). Further, by
environments’, the student questionnaire of TIMSS 2003 was standardising the regression coefficients we calculated effect sizes
extended with 29 items about the perceptions of four dimensions to compare the magnitude of significant effects in the models.
in the learning environment (Schelfhout, Vandenberghe, & Van Finally, the percentage of explained variance on each level was
Damme, 2002; Schelfhout et al., 2006) (see Appendix 2). The items calculated using the formulas of Snijders and Bosker (1999). These
gauged the frequency to which the teacher applies certain teaching formulas take into account the number of units on each level. For
activities and were answered on a four-point Likert scale the calculation of the percentages of explained variance on each
(1 = Almost always to 4 = Never). The responses on positively level we used the mean number of units within a group, that is, 19
formulated items were recoded, so that high item scores were an students per class and two classes per school.
indication of a higher frequency of the particular teaching activity. We estimated the multilevel models in MLwiN software
A confirmatory factor analysis (CFA) was accomplished to (Rasbash, Steele, Browne, & Goldstein, 2009) using the ‘Iterative
investigate whether the established dimensionality in the model of Generalized Least Squares’ (IGLS) algorithm. Because some
Schelfhout (2003) and Schelfhout et al. (2006) fitted the data. The students skipped some items in the questionnaire, missing
CFA model was estimated with the software program Mplus with a values in the items of the dependent variables were handled by
robust maximum likelihood estimator, which takes into account means of the Expectation Maximization algorithm in SPSS (Stern,
non-normality and missingness in the data. The analysis confirmed 2009). The amount of missing values per item ranged between
the four-dimension structure, i.e., Motivate to exert learning effort 0.2 and 1.5%. The calculation of the factor scores for each
(motivate), Activate towards self-regulated learning (activate), Give component of the mathematics attitude was based on the
feedback and coach (feedback) and Structure and steer (structure). imputed dataset. Missing values in the predictors were handled
An acceptable model fit was found (x2(435) = 42,531.37; by the IGLS algorithm in MLwiN. An overview of the number of
RMSEA = .06; SRMR = .05) (Schermelleh-Engel & Moosbrugger, values per variable, as well as their descriptive statistics, is
2003). The internal consistency (Cronbach’s a) for the scales shown in Appendix 3.
‘motivate’, ‘activate’, ‘feedback’ and ‘structure’ were .76, .83, .78
and .77 respectively. Intra class correlations ranged between .20 Results
and .27, indicating that students within classes shared consider-
able amounts of variance in their judgments. All between-class An unconditional or null model for each of the three
variances significantly differed from zero (p < .001). An overview components of the mathematics attitude was specified. The
of the questionnaire items for each factor is shown in Appendix 2. results are shown in Tables 1–3. For each of the three components,
The student level factor scores were calculated and for each between class and between school differences were particularly
112
Table 1
Model estimates of the multilevel analysis of the Mathematics academic self-concept.

Fixed effects Null model Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 ES

Effect SE Effect SE Effect SE Effect SE Effect SE Effect SE Effect SE

Intercept 0.002 0.023 0.004 0.066 0.120 0.066 0.119 0.065 0.119 0.065 0.124 0.065 0.124 0.065
Student level predictors
Cognitive abilities 0.039 0.001 0.041 0.001 0.041 0.001 0.041 0.001 0.041 0.001 0.041 0.001 0.47
Parental education level 0.034 0.019 0.039 0.019 0.039 0.019 0.039 0.019 0.039 0.019 0.039 0.019 0.03
Home possessions 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.01

M. Vandecandelaere et al. / Studies in Educational Evaluation 38 (2012) 107–120


Boy 0.306 0.030 0.315 0.029 0.315 0.029 0.314 0.029 0.313 0.029 0.314 0.029 0.16
Language 0.066 0.052 0.048 0.052 0.046 0.052 0.044 0.052 0.044 0.052 0.045 0.052 0.01
1st generation immigrant 0.146 0.080 0.120 0.080 0.118 0.080 0.117 0.080 0.117 0.080 0.117 0.080 0.07
2nd generation immigrant 0.071 0.049 0.055 0.044 0.056 0.049 0.056 0.049 0.055 0.049 0.056 0.049 0.03
General subject 0.181 0.043 0.296 0.045 0.301 0.045 0.302 0.045 0.304 0.045 0.303 0.045 0.22
Technical subject 0.013 0.056 0.301 0.067 0.306 0.066 0.311 0.066 0.312 0.067 0.311 0.067 0.23

Class level predictors


Mean cognitive abilities 0.030 0.004 0.030 0.004 0.030 0.004 0.03 0.004 0.030 0.004 0.00
Motivate 0.102 0.055 0.03
Activate 0.163 0.070 0.04
Feedback 0.398 0.207 0.03
Structure 0.097 0.057 0.03

Random effects
School level 0.016 0.010 0.036 0.009 0.020 0.007 0.02 0.007 0.02 0.007 0.02 0.007 0.02 0.007
Class level 0.036 0.012 0.012 0.007 0.009 0.007 0.009 0.007 0.008 0.007 0.008 0.007 0.009 0.007
Student level 0.948 0.021 0.753 0.017 0.752 0.017 0.752 0.017 0.752 0.017 0.752 0.017 0.752 0.017

Model fit
Deviance 12,281.547 10,737.841 10,692.878 10,689.459 10,687.5 10,689.210 10,689.968
x2 1543.733 44.936 3.419 5.658 3.666 2.91
df 9 1 1 1 1 1
p< .001 .001 .10 .05 .10 .10

ICC R2 R2 R2 R2 R2 R2

School level 1.6% – 24.9% 24.9% 25.7% 25.7% 24.9%


Class level 3.6% 14.0% 32.7% 32.7% 33.7% 33.7% 32.7%
Student level 94.8% 19.9% 21.9% 21.9% 22.0% 22.0% 21.9%

Note: Significant estimates and effect sizes >.10 are bolded. SE = standard error; ICC = intra class correlation; R2 = percentage of explained variance on a particular level; ES = effect size.
Table 2
Model estimates of the multilevel analysis of the Enjoyment in mathematics.

Fixed effects Null model Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 ES

Effect SE Effect SE Effect SE Effect SE Effect SE Effect SE Effect SE

Intercept 0.001 0.033 0.201 0.081 0.177 0.085 0.177 0.07 0.168 0.074 0.135 0.073 0.13 0.073

Student level predictors


Cognitive abilities 0.003 0.001 0.003 0.002 0.003 0.002 0.003 0.002 0.003 0.002 0.003 0.002 0.03

M. Vandecandelaere et al. / Studies in Educational Evaluation 38 (2012) 107–120


Parental education level 0.021 0.021 0.022 0.021 0.017 0.02 0.019 0.02 0.022 0.02 0.021 0.02 0.02
Home possessions 0.004 0.001 0.004 0.001 0.004 0.001 0.004 0.001 0.004 0.001 0.003 0.001 0.03
Boy 0.154 0.033 0.155 0.033 0.16 0.031 0.149 0.032 0.142 0.032 0.142 0.032 0.07
Language 0.025 0.057 0.026 0.057 0.022 0.055 0.031 0.056 0.035 0.056 0.034 0.056 0.01
1st generation immigrant 0.067 0.086 0.063 0.086 0.057 0.084 0.059 0.085 0.054 0.085 0.055 0.085 0.03
2nd generation immigrant 0.056 0.053 0.053 0.053 0.069 0.052 0.061 0.052 0.055 0.052 0.058 0.052 0.04
General subject 0.119 0.065 0.097 0.069 0.046 0.047 0.054 0.053 0.031 0.052 0.026 0.052 0.02
Technical subject 0.134 0.079 0.081 0.098 0.078 0.071 0.04 0.079 0.022 0.078 0.026 0.077 0.02

Class level predictors


Mean cognitive abilities 0.006 0.006 0.006 0.004 0.008 0.005 0.007 0.005 0.005 0.005 0.00
Motivate 1.072 0.058 0.34
Activate 1.212 0.086 0.30
Feedback 3.619 0.249 0.31
Structure 1.003 0.067 0.31

Random effects
School level 0.053 0.020 0.037 0.019 0.037 0.019 0.029 0.008 0.039 0.011 0.031 0.01 0.031 0.01
Class level 0.103 0.020 0.112 0.022 0.112 0.022 0.006 0.007 0.023 0.009 0.024 0.01 0.021 0.009
Student level 0.847 0.019 0.843 0.019 0.842 0.019 0.843 0.019 0.843 0.019 0.844 0.019 0.844 0.019

Model fit
Deviance 11,962.062 11,369.537 11,368.731 11,168.039 11,230.581 11,223.728 11,217.391
x2 592.525 0.806 200.692 138.15 145.003 151.34
df 9 1 1 1 1 1
p< .001 .5 .001 .001 .001 .001

ICC R2 R2 R2 R2 R2 R2

School level 5.3% 9.2% 9.2% 57.3% 42.7% 48.6% 49.8%


Class level 10.3% 3.7% 3.7% 60.4% 47.0% 50.4% 51.9%
Student level 84.4% 1.4% 1.5% 12.5% 9.8% 10.4% 10.7%

Note: Significant estimates and effect sizes >.10 are bolded. SE = standard error; ICC = intra class correlation; R2 = percentage of explained variance on a particular level; ES = effect size.

113
114
Table 3
Model estimates of the multilevel analysis of the Perceived value of mathematics.

Fixed effects Null model Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 ES

Effect SE Effect SE Effect SE Effect SE Effect SE Effect SE Effect SE

Intercept 0.003 0.022 0.032 0.070 0.081 0.073 0.080 0.072 0.081 0.072 0.087 0.072 0.087 0.072

Student level predictors


Cognitive abilities 0.003 0.002 0.002 0.002 0.002 0.002 0.002 0.002 0.002 0.002 0.002 0.002 0.02

M. Vandecandelaere et al. / Studies in Educational Evaluation 38 (2012) 107–120


Parental education level 0.045 0.021 0.043 0.021 0.043 0.021 0.043 0.021 0.042 0.021 0.043 0.021 0.03
Home possessions 0.005 0.002 0.005 0.002 0.005 0.002 0.005 0.002 0.005 0.002 0.005 0.002 0.05
Boy 0.130 0.032 0.125 0.033 0.125 0.032 0.127 0.032 0.128 0.032 0.128 0.032 0.06
Language 0.019 0.057 0.013 0.057 0.01 0.057 0.008 0.057 0.007 0.057 0.007 0.057 0.00
1st generation immigrant 0.077 0.088 0.094 0.088 0.098 0.088 0.098 0.088 0.099 0.088 0.099 0.088 0.06
2nd generation immigrant 0.018 0.054 0.028 0.054 0.027 0.054 0.028 0.054 0.029 0.054 0.028 0.054 0.02
General subject 0.005 0.049 0.050 0.052 0.058 0.052 0.058 0.052 0.061 0.053 0.062 0.052 0.05
Technical subject 0.154 0.059 0.044 0.073 0.038 0.073 0.034 0.073 0.029 0.073 0.029 0.073 0.02

Class level predictors


Mean cognitive abilities 0.011 0.005 0.012 0.004 0.012 0.004 0.012 0.004 0.011 0.004 0.00
Motivate 0.142 0.062 0.04
Activate 0.190 0.079 0.05
Feedback 0.535 0.232 0.04
Structure 0.151 0.063 0.05

Random effects
School level 0.010 0.010 0.008 0.009 0.008 0.008 0.007 0.008 0.006 0.008 0.007 0.008 0.007 0.008
Class level 0.041 0.012 0.029 0.011 0.028 0.011 0.027 0.011 0.028 0.011 0.028 0.011 0.028 0.011
Student level 0.949 0.021 0.928 0.021 0.928 0.021 0.927 0.021 0.927 0.021 0.927 0.021 0.927 0.021

Model fit
Deviance 12,285.791 11,575.992 11,569.904 11,564.770 11,564.130 11,564.660 11,564.260
x2 709.799 6.088 5.133 5.774 5.240 5.644
df 9 1 1 1 1 1
p< .001 .05 .05 .05 .05 .05

ICC R2 R2 R2 R2 R2 R2

School level 1.0% 15.4% 16.3% 19.1% 20.0% 18.2% 18.2%


Class level 4.1% 15.0% 16.0% 18.0% 18.0% 17.0% 17.0%
Student level 94.9% 3.5% 3.6% 3.9% 3.9% 3.8% 3.8%

Note: Significant estimates and effect sizes >.10 are bolded. SE = standard error; ICC = intra class correlation; R2 = percentage of explained variance on a particular level; ES = effect size.
M. Vandecandelaere et al. / Studies in Educational Evaluation 38 (2012) 107–120 115

small. The between school differences ranged between 1.0% and self-regulated learning, gave feedback and coached, and structured
5.3% and the between class differences ranged between 3.6% and and steered, the more students enjoyed mathematics. The effect
10.3%. These small percentages indicate that the majority of the sizes indicate a large impact of the dimensions in the learning
variance in mathematics attitude was situated between students environment on the enjoyment of mathematics. This is also
within classes, ranging from 84.4% for the enjoyment of apparent from the explained variance in the enjoyment of
mathematics to 94.9% for the perceived value of mathematics. mathematics. Whereas the addition of student characteristics
We report on the results of the six models, for each component of and the mean cognitive abilities only explained 9.2%, 3.7% and 1.4%
the mathematics attitude. of the variance at the school, class and student level respectively,
the addition of the dimensions of the learning environment
Mathematics academic self-concept increased the explained variance to a large extent. More
specifically, the additional explained school level variance ranged
Starting from the unconditional null model, we added between 33.5% in Model 4 and 48.1% in Model 3. At the class level,
explanatory variables. First, we included the seven student the additional proportion of explained variance ranged between
characteristics to examine their relation with the mathematics 43.3% in Model 4 and 56.7% in Model 3.
academic self-concept (see Table 1). Compared to the null model,
the inclusion of the student characteristics resulted in a significant Perceived value of mathematics
model improvement (x2(9) = 1543.733, p < .001). Next, the mean
cognitive abilities of the class group were included in Model 2, also First, the student characteristics were incorporated in the
resulting in a significant model improvement (x2(1) = 44.936, model, resulting in a significant improvement of the model
p < .001). In Models 3–6, the four scales concerning the learning (x2(9) = 709.799, p < .001). Second, the mean cognitive abilities of
environment were added, each separately. Only the addition the class group were included. This likewise improved the model
of ‘activate’ resulted in a significant model improvement fit (x2(1) = 6.088, p < .05). In the third to sixth model, each of the
(x2(1) = 5.658, p < .05). The addition of ‘motivate’, ‘feedback’ or four variables concerning the learning environment were added
‘structure’ did not result in a significant model improvement separately, which resulted in a significant model improvement
(2.910  x2(1)  3.666, p > .05). after adopting the scales ‘motivate’ (x2(1) = 5.133, p < .05),
The results reveal a significant effect of cognitive abilities, ‘activate’ (x2(1) = 5.774, p < .05), ‘feedback’ (x2(1) = 5.240,
parental education level, gender, and study section on the p < .05) and ‘structure (x2(1) = 5.644, p < .05).
mathematics academic self-concept. More specifically it appears The results show that gender and the two SES indicators (i.e.,
that more intelligent students and students with parents with a home possessions and parental education level) had a significant
higher education level reported a higher mathematics academic effect on the perceived value of mathematics, indicating that boys,
self-concept, as well as boys and students in a classical study students with more home possessions and students with higher
section compared to girls and students in a general or technical educated parents perceived mathematics as more valuable.
study section. Further, we found that the higher the mean Further, a significant effect of the mean cognitive abilities of the
cognitive abilities of the class group were, the lower the class group was found. The higher the mean cognitive abilities of
mathematics academic self-concept was. This result describes the class group were, the more valuable mathematics was
the Big-fish-little-pond effect. Concerning the learning environ- perceived. Finally, the four dimensions of the learning environ-
ment, a higher score on ‘activate’ indicated a lower academic self- ment were significantly and positively associated with the
concept. However, the effect size was small. The effect sizes perceived value of mathematics. Nevertheless, all effect sizes
indicate a large effect of the cognitive abilities of the student and were small. As Table 3 shows, the proportion explained variance in
the study section. As can be seen from Table 1, Models 1 and 2 the perceived value of mathematics increased mainly with the
explained most variance in the mathematics academic self- addition of the student characteristics, which explained 15.4% of
concept. More specifically, the inclusion of student characteristics the school level variance, 15.0% of the class level variance and
in Model 1 explained 19.9% of the student level variance and 14.0% merely 3.5% of the student level variance.
of the class level variance. In Model 2, the addition of the mean
cognitive abilities of the class group explained 24.9% of the Discussion
variance at the school level and 18.7% more of the variance at the
class level, compared to Model 2. The main objective of the present study was to gain insight
into the way and the degree to which teachers can stimulate a
Enjoyment in mathematics more positive attitude towards mathematics with their students.
The focus was on the potential impact of four dimensions in the
Model 1, which incorporated the effect of the student learning environment as presented in the ‘Equilibrium model for
characteristics on mathematics enjoyment, resulted in a significant creating powerful learning environments’, which are the extent
model improvement compared to the null model (x2(9) = 592.525, to which the teacher ‘motivates to exert learning’, ‘activates
p < .001). Adding the mean cognitive abilities of the class group towards self-regulated learning’, ‘gives feedback and coaches’
(Model 2) did not lead to a better model fit compared to Model 1 and ‘structures and steers’. These dimensions reflect teaching
(x2(1) = 0.806, p = .369). Finally, the four variables concerning the activities which are considered as indicators of effective
learning environment were included in Models 3–6, showing all a teaching. The association was studied with three components
significant improvement of the model fit compared to Model 2 of the mathematics attitude: ‘the mathematics academic self-
(138.150  x2(1)  200.692, p < .001). concept’, ‘the enjoyment of mathematics’ and ‘the perceived
Gender and home possessions were significant predictors. Boys value of mathematics’. To obtain an accurate measure of the
enjoyed mathematics less than girls and the score for home impact of the learning environment, we accounted for non-
possessions was negatively associated with enjoyment in mathe- randomness in class enrolment by including relevant background
matics. However, the effect sizes at the student level were characteristics.
particularly small. Interestingly, the results reveal a significant Results reveal that most of the variance in the three
association with each dimension in the learning environment. The components of the mathematics attitude is situated at the student
more the teacher motivated to exert learning, activated towards level. Small percentages on the group level are generally found in
116 M. Vandecandelaere et al. / Studies in Educational Evaluation 38 (2012) 107–120

studies on non-cognitive outcomes, indicating that the majority of educational activities which affect both the enjoyment of, and
the variance in the mathematics attitude is situated between beliefs about mathematics and contribute to the overall attitude
students within classes (Van Damme et al., 2004). This finding towards mathematics.
suggests that the attitude towards mathematics is mainly
associated with student characteristics, and to a lesser extent The importance of intake characteristics
with class characteristics. Nevertheless, the fact that we did find
differences between classes suggests a potential role of the It is worth noting that the findings support previous work on
learning environment in enhancing the attitude towards mathe- the effect of student characteristics on the mathematics attitude.
matics. In this section, the student and class variables with both a Boys reported higher scores on mathematics academic self-
significant association with the attitude component and an effect concept compared to girls, which is in line with other research
size larger than .10 are discussed, as effect sizes smaller than .10 (Chouinard et al., 1999; Frost et al., 1994; Leder, 1995; Meelissen
can be interpreted as trivial (Cohen, 1988). & Doornekamp, 2004; Van Damme et al., 2004). Students with
higher cognitive abilities also had a higher mathematics academic
Learning environment and students’ mathematics attitude self-concept. This likewise extends the conclusions from previous
research (Ma, 1997; Van Damme et al., 2004). Furthermore, the
The results suggest that the learning environment is associated results are in agreement with the findings of Van Damme et al.
with the enjoyment of mathematics. Students enjoy mathematics (2004) on the effect of the study section. Students in the general
to a higher extent when the teacher motivates to exert learning, and technical study section had a lower mathematics academic
activates towards self-regulated learning, provides feedback and self-concept compared to students in the classical study section.
coaches, and structures and steers. This finding underlines the With regard to the mean cognitive abilities of the class group, the
importance of teaching activities within each dimension of the results corroborate the Big-fish-little-pond effect (Marsh & Hau,
learning environment for students’ enjoyment in mathematics. 2003; Marsh et al., 2007). High mean cognitive abilities of the
Further, no association was found between the dimensions in the class group had a negative effect on the mathematics academic
learning environment and neither the mathematics academic self- self-concept, taking into account individual cognitive abilities.
concept, nor the perceived value of mathematics. Summarising, Marsh et al. (2007) state that equally able students have higher
while effective teaching, in terms of the ‘Equilibrium model’, is academic self-concepts in groups where the average abilities are
important for students’ enjoyment in mathematics, it does not low. The results of our study support this phenomenon,
affect student’s mathematics academic self-concept, neither their specifically in the domain of mathematics. The Big-fish-little-
perceived value of mathematics. pond effect found in our study suggests that students with less
Our results are in line with, yet nuance, other work in the field cognitive abilities are better off in a homogeneous group while
on the relation between the learning environment and the attitude students with more cognitive abilities are better off in a
towards mathematics. Whereas previous studies mostly addressed heterogeneous group, at least with regard to their mathematics
the association with an undifferentiated conceptualisation of academic self-concept.
mathematics attitude or focussed on the association with a single
dimension (e.g., Papanastasiou, 2008; Simonson & Maushak, 2001; Limitations and suggestions for further research
Van Damme et al., 2004), we addressed the association with three
components of the mathematics attitude. Recalling the general It is worth mentioning that the findings should be interpreted in
dimensional structure of attitude (Ajzen, 1988), our study suggests the context of some study limitations. The present study provides
that the learning environment is associated with students’ affective additional evidence with respect to the role of the learning
responses but not with the cognitive dimension of the mathematics environment for the mathematics attitude. It turned out that
attitude, in our study represented by the academic self-concept indicators of effective teaching, as presented in the ‘Four
towards mathematics and the perceived value of mathematics. A dimensions equilibrium model’, play a significant role for the
possible explanation for this might be that the enjoyment of enjoyment of mathematics, whilst the mathematics academic self-
mathematics is more sensitive to stimuli in the learning concept and the perceived value of mathematics are insensitive to
environment and can be enhanced in a relatively short period, this. Further research is needed to establish a greater degree of
while the cognitive dimension of attitude is more resistant to accuracy on this matter. It would be interesting to assess
change. Townsend and Wilton (2003) claim that students’ beliefs interaction effects between student variables and cross-level
about mathematics are formed by a history of experiences. interaction effects between student and class variables. For
Changing these beliefs is not something you do at the drop of a example, students with certain characteristics may benefit more
hat. Nevertheless, as the three attitudinal dimensions share a or less from aspects in the learning environment with regard to
synergistic relation (Maio & Haddock, 2010), it seems possible that their mathematics attitude. Differential effects of teaching
consistent application of teaching activities in the ‘Equilibrium activities within each dimension of the learning environment
model’ changes the cognitive, and subsequently the behavioural may also be at play. Additionally, associations with the perceived
dimension of mathematics attitude in the long run as well. Indeed, value of mathematics may be further explored, as we found no
as concluded by Forgas (2004), ‘‘affective states can have a highly substantial effects of student characteristics, nor of the learning
significant influence on the way people form, maintain, and change environment. Further research is strongly recommended to
their attitudes, and how attitudes and social information are identify teaching activities which simultaneously influence the
cognitively represented and categorized’’ (p. 149). On the other cognitive and affective dimension of mathematics attitude.
hand, changing negative attitudes might ask for a more compre- A limitation of this study follows from its cross-sectional
hensive approach than merely engaging in ‘effective teaching’, in nature. The results refer to correlations between variables. Hence,
terms of the ‘Equilibrium model’. Actions with elements affecting no conclusions can be drawn on the causality of effects. For
both cognition and emotion are more promising to have a higher instance, it is likely that the perceptions of the learning
probability of success. The results of the present study suggest that environment are not merely antecedents of the enjoyment of
teachers who specifically intend to increase the enjoyment of mathematics. The enjoyment of mathematics might affect the way
mathematics should adjust the learning environment with respect in which students perceive the learning environment as well.
to the ‘Equilibrium model’. Further research is necessary to identify Indeed, it is believed that attitudes shape people’s perceptions of
M. Vandecandelaere et al. / Studies in Educational Evaluation 38 (2012) 107–120 117

the social and physical world (Walther & Langer, 2004). Our data the findings in the present study can inform and inspire future
indicated that students within classes shared considerable experimental-type research.
amounts of variance in their perceptions of the learning Finally, we suggest to replicate the study in order to investigate
environment. However, our study does not rule out that the generalisability of our findings to students in other countries.
perceptions might be partially influenced by the students’ attitude.
Large randomised controlled trials could provide information for a
better understanding of the causality of the effects. We hope that Appendix A

Appendix 1
Attitude towards mathematics: overview of the questionnaire items for each component.

Item Loading Questionnaire Direction

Mathematics academic self-concept (a = .92)


In general, I’m good at mathematics. 0.84 F +
I’m pretty good at mathematics. 0.84 F +
I usually do well in mathematics. 0.79 I +
Not everyone can be gifted for every school subject. For mathematics, I’m not really gifted. 0.79 F
For some reason, I can’t master mathematics. 0.77 F
Mathematics is not one of my strengths. 0.75 I
Mathematics is more difficult for me than for many of my classmates. 0.72 I
I learn things quickly in mathematics. 0.71 I +
I would like a job that involved using mathematics. 0.51 I +
Sometimes, when I do not initially understand a new topic in mathematics, 0.41 I
I know that I will never really understand.

Enjoyment of mathematics (a = .93)


Mathematics is boring. 0.78 F
I find mathematics a pleasant school subject. 0.75 F +
Our lessons in mathematics are mostly fascinating and interesting. 0.75 F +
I’m sick of mathematics. 0.74 F
Especially for mathematics, I’m happy when class is over. 0.73 F
Without mathematics, school would be much more fun. 0.72 F
I enjoy learning mathematics. 0.66 I +
I would like to take more mathematics in school. 0.58 I +

Perceived value of mathematics (a = .83)


I think for most occupations, mathematics is not useful. 0.74 F
I believe mathematics has little use. 0.74 F
Most of mathematics can be useful later on. 0.74 F +
I think learning mathematics will help me in my daily life. 0.68 I +
For a lot of things, occurring daily, mathematics is useful. 0.63 F +
I need mathematics to learn other school subjects. 0.49 I +
I need to do well in mathematics to get the job I want. 0.42 I +
I need to do well in mathematics to get into the university of my choice. 0.37 I +
To be good at mathematics, is a case of luck. 0.32 F

Note: I = International part of TIMSS 2003; F = Flemish part of TIMSS 2003.


Responses on positively formulated items were reversed in order that a high score on an item was an indicator of a positive mathematics attitude.
118 M. Vandecandelaere et al. / Studies in Educational Evaluation 38 (2012) 107–120

Appendix 2
Perceptions of the learning environment: overview of the questionnaire items for each factor.

Item Loading

Motivate to exert learning effort (a = .76)


In the math class. . .
. . . the teacher makes sure that I get interested in the subject matter. 0.74
. . . the teacher uses an agreeable diversity of approaches in his/her teaching. 0.73
. . . we work in a pleasant manner. 0.73
. . . I feel that the subject matter will be useful to me later. 0.49

Activate towards self-regulated learning (a = .83)


In the math class. . .
. . . the teacher gives tasks that encourage us to keep looking for a solution. 0.60
. . . when we start with a new subject, the teacher takes time to repeat previous subject matter that will be relevant to the new topic. 0.59
. . . the teacher bears in mind pupils’ remarks when searching for suitable assignments or practice materials. 0.59
. . . each new chapter starts with examples from daily life that clarify the new subject. 0.56
. . . the teacher gives small clues that help us to find solutions by ourselves. 0.56
. . . situations are described that can happen in the real world and that need a mathematical solution. 0.55
. . . the teacher points out connections between new and previously treated subject matter. 0.55
. . . the teacher asks about relationships between different parts of the subject material during tasks. 0.55
. . . during team work or when I am working on my own, the teacher inquires after the difficulties I encounter while solving a problem. 0.53
. . . the teacher gives examples of situations in daily life where the subject matter can be applied. 0.52
. . . the teachers’ tests require us to apply the subject matter to new problem contexts. 0.44
. . . during team work or when I am working on my own, the teacher inquires after the time I need to solve a problem. 0.40

Give feedback and coach (a = .78)


In the math class. . .
. . . the teacher takes into account pupils’ answers. 0.71
. . . the teacher repeats the subject matter when it is not properly understood by some pupils. 0.66
. . . the teacher clarifies errors in tests. 0.64
. . . the teacher explains the solution after an exercise. 0.62
. . . when an assignment or test goes rather wrong, I am encouraged to think about what has caused the problem and what I can do about it. 0.59
. . . we have the opportunity to ask other pupils to explain their way of solving a problem. 0.55
. . . we have the opportunity to discuss our approach to math problems with other pupils. 0.50
. . . after a test we receive the solutions on paper. 0.17

Structure and steer (a = .77)


In the math class. . .
. . . it’s thanks to the teacher’s approach that I understand the subject matter well. 0.71
. . . the teacher is able to explain new topics in a clear and well-organised manner. 0.70
. . . we get the opportunity to explain our solution to the teacher. 0.63
. . . the teacher tries to make us understand new subject matter by alternating questions to the class with explanations. 0.63
. . . the teacher keeps the class under control. 0.47

Appendix 3
Description of variables.

Variable name N % M SD Min Max

Mathematics attitude
Mathematics academic self-concept 4354 0.00 1 2.83 2.72
Enjoyment in mathematics 4354 0.00 1 3.16 2.83
Perceived value of mathematics 4354 0.00 1 4.05 2.61

Student level predictors


Cognitive abilities 4334 52.31 11.48 9 80
Parental education level 4171 2.19 0.78 1 3
Home possessions 4353 77.28 10.11 6.25 100
Gender 4354
Female (0) 53.9
Male (1) 46.1
Language 4352
(Almost) never Dutch (0) 9.6
(Almost) always Dutch (1) 90.3
Origin 4348
Belgian (1) 86.7
First generation immigrant (2) 3.7
Second generation immigrant (3) 9.6
Study section 4354
Classical languages 22.8
General subjects 45.8
Technical options 31.4

Class level predictors


Mean cognitive abilities 228 52.30 6.49 35.40 67.60
Motivate 228 0.00 0.31 0.92 0.90
Activate 228 0.00 0.25 0.90 0.53
Feedback 228 0.00 0.08 0.33 0.17
Structure 228 0.00 0.31 1.22 0.59

Note: Coding of variables attributes is indicated by number in parentheses. N = number of observations; M = mean; SD = standard deviation; Min = minimum; Max = maximum.
M. Vandecandelaere et al. / Studies in Educational Evaluation 38 (2012) 107–120 119

References Kadijevich, D. (2006). Developing trustworthy TIMSS background measures: A case


study on mathematics attitude. The Teaching of Mathematics, 2, 41–51, Retrieved
Ajzen, I. (1988). Attitudes, personality and behavior. Milton Keynes: Open University from http://elib.mi.sanu.ac.rs.
Press. Kadijevich, D. (2008). TIMSS 2003: Relating dimensions of mathematics attitude to
Allen, E., & Seaman, C. (2007). Likert scales and data analyses. Quality Progress, 40(7), mathematics achievement (MA). Retrieved June, 14, 2009 from: http://www.doi-
64–65. serbia.nb.rs.
Armstrong, R. (1987). The midpoint on a five-point Likert-type scale. Perceptual and Kennedy, E., & Mandeville, G. (2000). Some methodological issues in school effective-
Motor Skills, 40, 359–362. ness research. In C. Teddlie & D. Reynolds (Eds.), The international handbook of
Bennet, R. E., & Ward, W. C. (1993). Construction versus choice in cognitive measurements. school effectiveness research. New York: Falmer Press.
Hillsdale: Erlbaum. Kogce, D., Yildiz, C., Aydin, M., & Altindag, R. (2009). Examining elementary school
Biemans, H. J. A. (1997). Fostering activation of prior knowledge and conceptual change students’ attitudes towards mathematics in terms of some variables. World
(Doctoral dissertation). Katholieke Universiteit Nijmegen, Nijmegen. Available conference on educational sciences - new trends and issues in educational
from http://opc.ubn.ru.nl/. sciences. Procedia Social and Behavioral Sciences Amsterdam: Elsevier Science BV
Boekaerts, M. (2001). Pro-active coping: Meeting challenges and achieving goals. In E. [retrieved from WOS:000275580400049].
Frydenberg (Ed.), Beyond coping: Meeting goals, visions and challenges. Oxford: Leder, G. (1995). Equity inside the mathematics classroom: Fact or artifact? In W. G.
Oxford University Press. Secada, E. Fennema, & L. B. Adajian (Eds.), New directions for equity in mathematics
Bonwell, C. C., & Eison, J. A. (1992). Active learning: Creating excitement in the classroom. education (pp. 209–224). New York: Cambridge University Press.
ERIC Digest. Retrieved from ERIC database. (ED340272). Ma, X. (1997). Reciprocal relationships between attitude toward mathematics and
Chouinard, R., Vezeau, C., Bouffard, T., & Jenkins, B. (1999). Gender differences in the achievement in mathematics. Journal of Educational Research, 90, 221–229, Re-
development of mathematics attitudes. Journal of Research and Development in trieved from http://www.jstor.org/stable/27542096.
Education, 32, 184–192 http://dx.doi.org/10.1037/0033-2909.107.2.139. Ma, X., & Kishore, N. (1997). Assessing the relationship between attitude toward
Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). New mathematics and achievement in mathematics: A meta-analysis. Journal for Re-
Jersey: Lawrence Erlbaum. search in Mathematics Education, 28, 26–47.
Corno, L., & Randi, J. (1997). Motivation, volition and collaborative innovation in Maio, G. R., & Haddock, G. (2010). The psychology of attitudes and attitude change.
classroom literacy. In J. Guthrie & A. Wigfield (Eds.), Reading engagement: Motivat- London: Sage.
ing readers through integrated instruction (pp. 14–31). Newark: International Read- Marks, G. (1999). The measurement of socioeconomic status and social class in the LSAY
ing Association. (ED404627). project. LSAY Publications, NCVER. Retrieved from http://www.lsay.edu.au/pub-
Cox, E. P. (1980). The optimal number of response alternatives for a scale: A review. lications/1931.html.
Marketing Research, 17, 407–422 http://dx.doi.org/10.2307/3150495. Marks, G., McMillan, J., Jones, F. L., & Ainley, J. (2000). The measurement of socioeconomic
Crombie, G., Sinclair, N., Silverthorn, N., Byrne, B. M., Dubois, D. I., & Trinneer, A. (2005). status for the reporting of nationally comparable outcomes of schooling. Report
Predictors of young adolescents’ math grades and course enrollment intentions: prepared for the National Education Performance Monitoring Taskforce. Retrieved
Gender similarities and differences. Sex Roles, 52(5–6), 351–367 http://dx.doi.org/ from http://www.mceetya.edu.au/verve/_resources/socioeconomicstatus_file.
10.1007/s11199-005-2678-1. pdf.
Daskalogianni, K., & Simpson, A. (2000). Towards a definition of attitude: The relation- Marsh, H. W., & Craven, R. (2006). Reciprocal effects of self-concept and performance
ship between the affective and the cognitive in pre-university students. In Naka- from a multidimensional perspective: Beyond seductive pleasure and unidimen-
hara, T., & Koyama, M. (Eds.), Proceedings of the 24th conference of the international sional perspectives. Perspectives on Psychological Science, 1, 133–163 http://
group for the psychology of mathematics education (Vol. 3, pp. 217–224).Hiroshima, dx.doi.org/10.1111/j.1745-6916.2006.00010.x.
Japan: PME . Marsh, H. W., & Hau, K. (2003). Big-fish-little-pond effect: A cross-cultural (26-country)
test of the negative effects of academically selective schools. American Psychologist,
De Jong, R., & Westerhof, K. J. (2001). The quality of student ratings of teacher
behaviour. Learning Environments Research, 4(1), 51–85 http://dx.doi.org/ 58(5), 364–376 http://dx.doi.org/10.1037/0003-066X.58.5.364.
10.1023/A:1011402608575. Marsh, H. W., Trautwein, U., Lüdtke, O., Baumert, J., & Köller, O. (2007). The Big-fish-
Demeyer, W. (1999). Niet-Verbale Testbatterij voor het Lager Secundair Onderwijs. De little-pond effect: Persistent negative effects of selective high schools on self-
gestandaardiseerde aanpassing en bewerking in het Nederlands van de B.C.R. (Batterie concept after graduation. American Educational Research Journal, 44, 631–669
de tests du Centre de Recherches). [Non-verbal test battery for the seventh and http://dx.doi.org/10.3102/0002831207306728.
eighth grade. A standardized adaptation and revision to Dutch]. Brussels: EDITEST. Martinot, M. J., Kuhlemeier, H. B., & Feenstra, H. J. M. (1988). Het meten van affectieve
Di Martino, P., & Zan, R. (2010). ‘Me and maths’: Towards a definition of attitude doelen: De validering en normering van de belevingsschaal voor wiskunde (BSW)
grounded on students’ narratives. Journal of Mathematics Teacher Education, 13, 27– [[Measuring affective goals: Validation and standardization of the Attitude Scale
48 http://dx.doi.org/10.1007/s10857-009-9134-z. towards Mathematics (ASM)]. Tijdschrift voor onderwijsresearch, 13, 65–76.
Di Martino, P., & Zan, R. (2011). Attitude towards mathematics: A bridge between Massey, D. S., & Denton, N. A. (1988). Suburbanization and segregation in U.S.
beliefs and emotions. ZDM Mathematics Education, 43, 471–482. metropolitan areas. American Journal of Sociology, 94, 592–626.
Forgas, J. (2004). The role of affect in attitudes and attitude change. In W. D. Crano & R. McDowell, L. (1995). The impact of innovative assessment on student learning.
Prislin (Eds.), Attitudes and attitude change (pp. 131–160). New York: Psychology Innovations in Education and Training International, 32, 302–313 http://
Press. dx.doi.org/10.1080/1355800950320402.
Foy, P., & Joncas, M. (2004). TIMSS 2003 sampling design. In M. Martin, I. Mullis, & S. Meelissen, M., & Doornekamp, B. (2004). TIMSS-2003 Nederland: Leerprestaties in exacte
Chrostowski (Eds.), TIMSS 2003 technical report (pp. 109–124). Boston: TIMSS & vakken in het voortgezet onderwijs. [TIMSS-2003 The Netherlands. Students’
PIRLS International Study Centre. Retrieved from http://timss.bc.edu/PDF/ achievement in mathematics and science in grade 8, secondary education].
t03_download/T03_TR_Chap5.pdf. Enschede: University of Twente. Retrieved from http://doc.utwente.nl/72736/1/
Frost, L. A., Hyde, J. S., & Fennema, E. (1994). Gender, mathematics performance, and TIMMS2003VO.pdf.
mathematics related attitudes and affect: A meta-analytic synthesis. International Mettas, A., Karmiotis, I., & Chirstoforou, P. (2006). Relationship between students’ self-
Journal of Educational Research, 21, 373–385 http://dx.doi.org/10.1016/S0883- beliefs and attitudes on science achievements in Cyprus: Findings from the Trends
0355(06)80026-1. in International Mathematics and Science Study (TIMSS). Eurasia Journal of Mathe-
Garaway, G. B. (1994). Language, culture, and attitude in mathematics and science matics, Science and Technology Education, 2, 41–52, Retrieved from http://
learning: A review of the literature. Journal of Research and Development in Educa- www.ejmste.com/012006/d3.pdf.
tion, 27(2), 102–111. Middleton, J. A., & Spanias, P. A. (1999). Motivation for achievement in mathematics:
Gómez-Chacón, I. M. (2000). Affective influences in the knowledge of mathematics. Findings, generalizations and criticisms of the research. Journal of Research in
Educational Studies in Mathematics, 43(2), 149–168. Mathematics Education, 30, 65–88 http://dx.doi.org/10.2307/749630.
Good, T., & Brophy, J. (2000). Looking in classrooms (8th ed.). New York: Longman. Mullis, I. V. S., Martin, M. O., & Stemler, S. E. (2000). TIMSS questionnaire development.
Groves, R. M., Floyd, J. F., Couper, M. P., Lepkowski, J. M., Singer, E., & Tourangeau, R. In M. O. Martin, K. D. Gregory, & S. E. Stemler (Eds.), TIMSS 1999 technical report (pp.
(2011). Survey methodology. John Wiley & Sons. 69–86). Chestnut Hill, MA: Boston College.
Hanulla, M. S. (2002). Attitude towards mathematics: Emotions, expectations and O’Brien, R. M. (2007). A caution regarding rules of thumb for variance inflation factors.
values. Educational Studies in Mathematics, 49, 25–46 http://dx.doi.org/10.1023/ Quality & Quantity, 41, 673–690.
A:1016048823497. Pahlavan, S. A., Farzad, V. E., & Naderi, E. (2007). The relation between the family socio-
Helmke, A. (1989). Incentive value of success and failure in school. Developmental economic status, student characteristics, and math achievement as mirrored in
trends and impact on academic achievement. In F. Halisch & J. J. L. van den Berchen TIMSS 2003 data. Journal of Education, 22(4), 33–55.
(Eds.), International perspectives on achievement and task motivation (pp. 225–237). Papanastasiou, C. (2008). A residual analysis of effective schools and effective teaching
Lisse: Swets & Zeitlinger. in mathematics. Studies in Educational Evaluation, 34, 24–30.
Hox, J. J. (2010). Multilevel analysis. Techniques and applications. New York/Hove: Prawat, R. S. (1989). Promoting access to knowledge, strategy and disposition in
Routledge. Retrieved from http://ebookee.org/Multilevel-Analysis-Techniques- students: A research synthesis. Review of Educational Research, 59, 1–41 http://
and-Applications_1060083.html. dx.doi.org/10.2307/1170445.
Ismail, N. A. (2009). Understanding the gap in mathematics achievement of Malaysian Randhawa, B. S., & Gupta, A. (2000). Cross-national gender differences in mathematics
students. The Journal of Educational Research, 102, 389–394 http://dx.doi.org/ achievement, attitude and self-efficacy within a common intrinsic structure.
10.3200/JOER.102.5.389-394. Canadian Journal of School Psychology, 15(2), 51–66 http://dx.doi.org/10.1177/
Jones, R. G. (2001). Identifying higher education students from low socio-economic status 08295735000150020.
backgrounds and regional and remote areas. Department of Education, Science and Rasbash, J., Steele, F., Browne, W., & Goldstein, H. (2009). A user’s guide to MLwiN version
Training, Government of Australia. Retrieved from http://www.dest.gov.au/ar- 2.10. London: Institute of Education. Retrieved from www.bristol.ac.uk/cmm/
chive/highered/eippubs/eip02_4/index.htm#Contents. software/mlwin/download/manual-print.pdf.
120 M. Vandecandelaere et al. / Studies in Educational Evaluation 38 (2012) 107–120

Raudenbush, S. W., & Willms, J. D. (1995). The estimation of school effects. Journal of Singh, K., Granville, M., & Dika, S. (2002). Mathematics and science achievement:
Educational and Behavioral Statistics, 20(4), 307–335. Effects of motivation, interest and academic engagement. The Journal of Educational
Reigeluth, C. M. (1999). The elaboration theory: Guidance for scope and sequence Research, 95, 323–332 http://dx.doi.org/10.1080/00220670209596607.
decisions. In C. M. Reigeluth (Ed.), Instructional design theories and models (pp. 335– Snijders, T., & Bosker, R. (1999). Multilevel analysis: An introduction to basic and advance
381). Hillsdale: Lawrence Erlbaum Associates. multilevel modeling. London: Sage Publications.
Reynders, T., Nicaise, I., & Van Damme, J. (2005). De constructie van een SES-variabele Stern, L. D. (2009). A visual approach to SPSS for Windows: A guide to SPSS 17.0. Boston:
voor het SiBO-onderzoek Retrieved from http://www.steunpuntloopbanen.be/rap- Allyn & Bacon.
porten/LOA-rapport_31.pdf. Stipek, D. (2002). Motivation to learn. Integrating theory and practice. Boston: Allyn and
Reynolds, A. J., & Walberg, H. J. (1992). A structural model of science achievement and Bacon.
attitude: An extension to high school. Journal of Educational Psychology, 84, 371– Tahar, N. F., Ismail, Z., Zamani, N. D., & Adnan, N. (2010). Students’ attitude toward
382, (EJ452407). mathematics: The use of factor analysis in determining the criteria. Amsterdam:
Ruffell, M., Mason, J., & Allen, B. (1998). Studying attitude to mathematics. Educational Elsevier Science BV.
Studies in Mathematics, 35, 1–18 http://dx.doi.org/10.1023/A:1003019020131. Tapia, M., & Marsh, G. E. (2005). An instrument to measure mathematics attitudes.
Ryan, R., & Deci, E. (2000). When rewards compete with nature: The undermining of Academic Exchange Quarterly, 8, 16–21, Retrieved from http://www.rapidintellect.-
intrinsic motivation and self-regulation. In C. Sansone & J. Harackiewicz (Eds.), com/AEQweb/cho25344l.htm.
Intrinsic and extrinsic motivation: The search for optimal motivation and performance Tesser, A. (1993). The importance of heritability in psychological-research – The case of
(pp. 13–54). San Diego: Academic Press http://dx.doi.org/10.1016/B978- attitudes. Psychological Review, 100, 129–142.
012619070-0/50024-6. Thorndike-Christ, T. (1991). Attitudes toward mathematics: Relationships to mathematics
Schelfhout, W. (2003). Educating for constructivist teaching. Determinant factors and achievement, gender, mathematics course-taking plans, and career interests. Belling-
successful learning environments in teacher education. Paper presented at the annual ham: Western Washington University. (ED347066).
conference of the European Association for Research on Learning and Instruction Tobias, S., & Weissbrod, C. (1980). Anxiety and mathematics: An update. Harvard
(EARLI), Padova, 28 August. Educational Review, 50, 63–70, (EJ224889).
Schelfhout, W., Dochy, F., Janssens, S., Struyven, K., & Gielen, S. (2006). Towards an Townsend, M., & Wilton, K. (2003). Evaluating change in attitude towards mathematics
equilibrium model for creating a powerful learning environment. Validation of a using the ‘then-now’ procedure in a cooperative learning programme. British
questionnaire on creating powerful learning environments during teacher training Journal of Educational Psychology, 73, 473–487.
internships. European Journal of Teacher Education, 29(4), 471–503 http:// Tymms, P. (2001). A test of the big fish in a little pond hypothesis: An investigation into
dx.doi.org/10.1080/02619760600944787. the feelings of seven-year-old pupils in school. School Effectiveness and School
Schelfhout, W., Van Landeghem, G., Van den Broeck, A., & Van Damme, J. (2007). The Improvement, 12(2), 161–181.
effect of students’ perceptions of the learning environment on mathematics Utman, C. (1997). Performance effects of motivational state: A meta-analysis. Person-
achievement: Explaining variance in Flemish TIMSS 2003 data. In Wagemaker, ality and Social Psychology Review, 1, 170–182 http://dx.doi.org/10.1207/
P. (Ed.). The second IEA international research conference: Proceedings of the IRC-2006 s15327957pspr0102_4.
(Vol. 1, pp. 229–240). Wellington: Excel Digital Print. Retrieved from https:// Van Damme, J., Opdenakker, M., & Van den Broeck, A. (2004). Do classes and schools
lirias.kuleuven.be/bitstream/123456789/165284/1/Schelfhout_et_all.pdf. have an effect on attitudes? In C. Papanastasiou (Ed.), TIMSS (Proceedings of the IEA
Schelfhout, W., Vandenberghe, N., & Van Damme, J. (2002). Rapport vooronderzoek International Research Conference (IRC-2004)) (pp. 9–15). Lefkosia: Cyprus Univer-
TIMSS-2003: onderzoek naar constructivisme en metacognitie. [Report preliminary sity Press. Retrieved from http://www.iea.nl/fileadmin/user_upload/IRC/
study TIMSS-2003: Research on constructivism and metacognition (Research IRC_2004/Papers/IRC2004_VanDamme_Opdenakker_etal.pdf.
Report)]. Leuven: Centre of Educational Effectiveness and Evaluation. Van Damme, J., Opdenakker, M.-C., Van Landeghem, G., De Fraine, B., Pustjens, H., & Van
Schermelleh-Engel, K., & Moosbrugger, H. (2003). Evaluating the fit of structural de gaer, E. (2006). Educational effectiveness. An introduction to international and
equation models: Tests of significance and descriptive goodness-of-fit measures. Flemish research on schools, teachers and classes. Leuven: Acco.
Methods of Psychological Research Online, 8(2), 23–74, Retrieved from http://user.- Vygotsky, L. S. (1978). Mind in society: The development of higher psychological processes.
uni-frankfurt.de/kscherm/schermelleh/mpr_Schermelleh.pdf. Cambridge: Harvard University Press.
Schreiber, J. B. (2000). Advanced mathematics achievement: A hierarchical Walther, E., & Langer, T. (2004). Attitude formation and change through association: An
linear model. Dissertation Abstracts International, 61(8), 3056A, (UMI No. evaluative conditioning account. In W. D. Crano & R. Prislin (Eds.), Attitudes and
9981039). attitude change (pp. 131–160). New York: Psychology Press.
Schunk, D. (2001). Social cognitive theory and self-regulated learning. In B. Zimmer- Watt, H. M. G. (2000). Measuring attitudinal change in mathematics and English over
man & D. Schunk (Eds.), Self-regulated learning and academic achievement (pp. 125– the first year of junior high school: A multidimensional analysis. Journal of
152). Mahwah, NJ: Lawrence Erlbaum Associates Publishers. Experimental Education, 68, 331–361, Retrieved from http://users.monash.e-
Seligman, M. E. P., & Csikszentmihalyi, M. (2000). Positive psychology: An introduc- du.au/hwatt/articles/Watt_JXE_2000.pdf.
tion. American Psychologist, 55, 5–14 http://dx.doi.org/10.1037/0003- Willms, J. D. (2002). Monitoring school performance: A guide for educators. Washington,
066X.55.1.5. DC: RoudledgeFalmer.
Simonson, M., & Maushak, N. (2001). Instructional technology and attitude change. In Winne, P. H. (1997). Experimenting to bootstrap self-regulated learning. Journal of
D. H. Jonassen (Ed.), Handbook of research for educational communications and Educational Psychology, 89(3), 397–410, (EJ553135).
technology. New York: Simon & Schuster Macmillan. Retrieved from http:// Zimbardo, P. G., & Leippe, M. R. (1991). The psychology of attitude change and social
www.aect.org/edtech/ed1/firstedition.asp. influence. McGraw-Hill.

You might also like