Professional Documents
Culture Documents
Surviving Your Dissertation: A Comprehensive Guide to Content and Process 4th Edition, (Ebook PDF) full chapter instant download
Surviving Your Dissertation: A Comprehensive Guide to Content and Process 4th Edition, (Ebook PDF) full chapter instant download
https://ebookmass.com/product/designing-and-conducting-survey-
research-a-comprehensive-guide-4th-edition-ebook-pdf/
https://ebookmass.com/product/everybody-writes-your-go-to-guide-
to-creating-ridiculously-good-content-2nd-edition-ann-handley/
https://ebookmass.com/product/a-comprehensive-guide-to-
intellectual-and-developmental-disabilities-2nd-edition-ebook-
pdf/
https://ebookmass.com/product/original-pdf-professional-
counseling-a-process-guide-to-helping-8th-edition/
(eTextbook PDF) for Engineering Your Future A
Comprehensive Introduction to Engineering 9th Edition
https://ebookmass.com/product/etextbook-pdf-for-engineering-your-
future-a-comprehensive-introduction-to-engineering-9th-edition/
https://ebookmass.com/product/lymphedema-management-the-
comprehensive-guide-for-practitioners-4th-edition-ebook-pdf/
https://ebookmass.com/product/questions-and-answers-a-guide-to-
fitness-and-wellness-4th-edition-ebook-pdf/
https://ebookmass.com/product/learning-in-adulthood-a-
comprehensive-guide-4e-4th-edition-sharan-b-merriam/
https://ebookmass.com/product/a-comprehensive-russian-
grammar-4th-edition-terence-wade/
We believe that this book is suitable for a large academic and professional audience. Of course, it is primarily
directed at the graduate student who envisions or is involved in writing a research dissertation. Thus, there is
significant focus on material that is best suited for the beginning doctoral student—for example, how to
develop a research question, how to construct a table or figure, how to report a statistical finding, how to use
American Psychological Association formatting conventions, and so on. However, there is also content
directed at the more advanced student—for example, how to conceptualize and illustrate a mediation model,
how to report multiple regression findings, and how to code text for a grounded theory study. Moreover, we
have become increasingly aware that Surviving Your Dissertation serves also as a resource for researchers and
practitioners who have either forgotten important details or are motivated to keep abreast of evolving research
practices in their fields. Perhaps more important, we view the book as a convenient source of information for
faculty who are currently supervising graduate students’ dissertations or research projects.
We remain deeply indebted to our own students, who continue to thrill us with their creativity and force us to
keep learning in order to stay a step ahead. We hope that they, and you, find this volume a helpful and steady
companion in your research and writing endeavors.
A large number of individuals have contributed to the completion of this project. We called upon many
faculty colleagues to nominate student dissertations that exemplify high levels of scholarship and have
sprinkled references to these dissertations throughout the book to illustrate important principles and
recommendations. We are very appreciative of these relatively recent graduates for allowing us to share their
first major research endeavors in this venue. We also benefitted from the critical reflections and insight of the
following reviewers of the third edition of Surviving Your Dissertation: Anne J. Hacker, Bernie Kerr, Karin
Klenke, Kaye Pepper, and Udaya R. Wagle. Their observations and suggestions were both reinforcing and
helpful in crystallizing changes for this edition.
We are also grateful to be part of the SAGE family, a collaborative, dedicated group of professionals who have
facilitated our writing careers in so many ways. The following individuals were notable contributors: Vicki
Knight, publisher and senior editor, has always provided us with a balance of inspiring leadership and
nurturing support. Her editorial assistant, Jessica Miller, has been consistently responsive to our frequent
requests for assistance. Laura Barrett and David Felts, project editors, have gracefully guided the editorial
process from start to finish. And Paula Fleming, our copy editor, is truly a paragon in her craft. Her
grammatical acuity, common sense, and work ethic cannot be overestimated. Thank you all.
Finally, we must thank our partners in life, Jan and Kathy, for their continuing patience and support as we
have devoted our energy and attention to four editions of this volume.
8
About the Authors
9
Part I Getting Started
10
1 The Research Process
There is a story about a Zen Buddhist who took a group of monks into the forest. The group soon lost its
way. Presently one of the monks asked the leader where they were going. The wise man answered, “To the
deepest, darkest part of the forest so that we can all find our way out together.” Doctoral research for the
graduate student in the social sciences is often just such an experience—trekking into a forest of impenetrable
density and making many wrong turns. Over the years, our students have used various metaphors to describe
the dissertation process, metaphors that convey the feeling of being lost in the wilderness. One student
compared the process to the Sisyphean struggle of reaching the top of a hill, only to discover the presence of
an even higher mountain behind it. Another student experienced the task as learning a Martian language,
known to the natives who composed her committee but entirely foreign to her. A third student offered
perhaps the best description when she suggested that it was like waiting patiently in a seemingly interminable
line to gain admission to a desirable event, then finally reaching the front only to be told to return to the rear
of the line.
One reason that students become more exasperated than necessary on the dissertation journey is that they fail
to understand the procedures and practices that form the foundation for contemporary social science research.
Many students who are attracted to their field of interest out of an applied concern are apprehensive about
making the leap from application to theory, an indispensable part of the research enterprise. What may not be
so evident is that many of the skills that go into being a consummate practitioner are the same ones demanded
of a capable researcher. It is well known that curiosity and hypothesis testing are the bedrock of empirical
research. In a similar fashion, experienced psychotherapists, to take an example from clinical psychology, are
sensitive and keen observers of client behavior. They are persistent hypothesis testers. They are curious about
the relationship between family history variables and current functioning. They draw on theory and experience
to help select a particular intervention for a particular client problem or moment in therapy.
Dispassionate logic and clear and organized thinking are as necessary for effectiveness in the field as they are
for success in research. In fact, the bridge between research and just plain living is much shorter than most
people think. All of us gather data about the world around us, wonder what will happen if we or others behave
in particular ways, and test our pet hunches through deliberate action. To a large extent, the formal research
enterprise consists of thinking systematically about these same issues.
The procedures outlined in this book are intended to assist the doctoral student in planning and writing a
research dissertation, but the suggestions are equally applicable to writing a master’s thesis. In fact, there is
considerable overlap between these two challenging activities. For most students, the master’s thesis is the first
rigorous research project they attempt. This means that, in the absence of strong, supportive faculty
consultation, the student often concludes the thesis with considerable relief and an awareness of how not to do
the study the next time! With a doctoral dissertation, it is generally expected, sometimes as an act of faith,
that the student is a more seasoned and sophisticated researcher. The consensus opinion is that dissertations
are generally longer than theses, that they are more original, that they rely more heavily on theoretically based
11
arguments, and that they make a greater contribution to the field.
In most graduate programs, the prelude to conducting a dissertation study is presenting a dissertation
proposal. A research proposal is an action plan that justifies and describes the proposed study. The completion
of a comprehensive proposal is a very important step in the dissertation process. The proposal serves as a
contract between the student and his or her dissertation or thesis committee that, when approved by all
parties, constitutes an agreement that data may be collected and the study may be completed. As long as the
student follows the steps outlined in the proposal, committee members should be discouraged from
demanding significant changes to the study after the proposal has been approved. Naturally, it is not
uncommon to expect small changes, additions, or deletions as the study progresses because one can never
totally envision the unpredictable turns that research can take.
There is no universally agreed-on format for the research proposal. To our way of thinking, a good proposal
contains a review of the relevant literature, a statement of the problem and the associated hypotheses, and a
clear delineation of the proposed method and plan for data analysis. In our experience, an approved proposal
means that a significant percentage of the work on the dissertation has been completed. As such, this book is
intended to help students construct research proposals as well as complete dissertations.
12
The Research Wheel
One way to think about the phases of the research process is with reference to the so-called research wheel
(see Figure 1.1). The wheel metaphor suggests that research is not linear but is rather a recursive cycle of
steps. The most common entry point is some form of empirical observation. In other words, the researcher
selects a topic from the infinite array of possible topics. The next step is a process of inductive logic that
culminates in a proposition. The inductive process serves to relate the specific topic to a broader context and
begins with some hunches in the form “I wonder if . . . .” These hunches typically are guided by the values,
assumptions, and goals of the researcher, which need to be explicated.
Stage 2 of the research wheel is a developed proposition, which is expressed as a statement of an established
relationship (e.g., “the early bird is more likely than the late bird to catch the worm”). The proposition exists
within a conceptual or theoretical framework. The role of the researcher is to clarify the relationship between
a particular proposition and the broader context of theory and previous research. This is probably the most
challenging and creative aspect of the dissertation process.
Theories and conceptual frameworks are developed to account for or describe abstract phenomena that occur
under similar conditions. A theory is the language that allows researchers to move from observation to
observation and make sense of similarities and differences. A conceptual framework, which is simply a less-
developed form of a theory, consists of statements that link abstract concepts (e.g., motivation, role) to
empirical data. If not placed within such a context, the proposed study has a “So what?” quality. This is one of
the main objections to the research proposals of novice researchers: The research question may be inherently
interesting but ultimately meaningless. For instance, the question “Are there more women than men in
graduate school today?” is entirely banal as a research question unless the answer has conceptual or theoretical
implications that are developed within the study.
Although a study may be worthwhile primarily for its practical implications (e.g., “Should we start recruiting
13
more men into graduate schools?”), a purely applied study may not be acceptable as a dissertation. Kerlinger
and Lee (1999), authors of a highly respected text on research methodology, noted that “the basic purpose of
scientific research is theory” (p. 5). Generally speaking, a research dissertation is expected to contribute to the
scholarly literature in a field and not merely solve an applied problem. Thus, developing a proposition for
one’s dissertation typically involves immersing oneself in the research and theoretical literature of the field to
identify a conceptual framework for the study.
Having stated our position on the role of theory in dissertation research, we now need to take a step back. As
a psychologist and a sociologist, respectively, we are most familiar with research conventions within these two
disciplines. Other branches of the social sciences have their own standards of what constitutes an acceptable
dissertation topic. We have attempted to keep this book as generalizable as possible and to infuse it with
examples from other fields. Ultimately, of course, you will need to follow the rules and conventions that
pertain to your discipline as well as to your university and department.
For example, a few major universities allow a doctoral student to submit one or more published articles as the
equivalent of a dissertation. Many others encourage studies that consist of secondary data analyses derived
from national databases, such as U.S. Census data or the General Social Survey, or data obtained from a larger
study. Some fields—notably social work, education, policy evaluation, and professional psychology—may
encourage dissertations that solve applied problems rather than make distinct theoretical contributions.
Studies that evaluate the effectiveness of programs or interventions are a case in point because they sometimes
contribute little to validating a theory. Political science and economics are examples of fields that are diverse
enough to accommodate both theoretically based studies and purely applied studies. Within the subspecialty
of international relations, for instance, one could imagine a survey and analysis of security agreements of
European nations after the unraveling of the North Atlantic Treaty Organization (NATO) that rely on
interviews with foreign policy makers and are largely descriptive and applied. In contrast, a study of the role of
a commitment to ideology in the success of political parties in the United States, based on an analysis of
historical documents and voting records, might be grounded in a theory of how ideology attracts or alienates
the voting public.
Moving forward along the research wheel, the researcher uses deductive reasoning to move from the larger
context of theory to generate a specific research question. The research question is the precisely stated form of
the researcher’s intent and may be accompanied by one or more specific hypotheses. The first loop is
completed as the researcher seeks to discover or collect the data that will serve to answer the research question.
The data collection process is essentially another task of empirical observation, which then initiates another
round of the research wheel. Generalizations are made on the basis of the particular data that have been
observed (inductive process), and the generalizations are tied to a conceptual framework, which then leads to
the elucidation of further research questions and implications for additional study.
The kinds of skills called for at the various points of the research wheel are reminiscent of the thoughts about
learning presented by Bertrand Russell many years ago. Russell noted that there are two primary kinds of
knowledge acquisition: knowledge by description and knowledge by acquaintance. Knowledge by description
14
is learning in a passive mode, such as by reading a book on how to change the oil in one’s car or hearing a
lecture on Adam Smith’s theory of economics. This type of learning is especially well suited to mastering
abstract information; in other words, it is better for learning about economics than about changing the oil.
Knowledge by acquaintance, on the other hand, is learning by doing—the kind of skill training that comes
from practicing a tennis serve, driving an automobile, and playing with a computer. This is concrete
knowledge acquisition, oriented to solving problems.
The research process demands both skills. First, the researcher needs to apply clear, logical thinking to
working with concepts and ideas and building theories. It is our impression that many graduate students,
particularly those who have experience as practitioners in their fields, are weaker in this abstract
conceptualization, and honing this skill may be the major challenge of the dissertation. Second, the researcher
must engage in the practical application of ideas, including by systematically planning a study and then
collecting and analyzing data. The ability to focus, problem solve, and make decisions will help bring the
study to completion.
15
2 Selecting a Suitable Topic
The selection of an appropriate topic is the first major challenge in conducting research. In many academic
settings, this task is simplified by working with a faculty mentor who is already familiar with an interesting
area of study, may have an extensive program of research in that area, and may even have defined one or more
researchable questions. It is quite common for students interested in a particular area of research to not only
select their doctoral institution but also select their dissertation chair, with the goal of joining the research
program of a noted scholar in that field.
On the other hand, you may not be blessed with a faculty role model who is actively engaged in research in an
area of interest to you. There are no simple rules for selecting a topic of interest, but there are some
considerations with respect to appropriateness. It is generally unwise to define something as important as a
dissertation topic without first obtaining a broad familiarity with the field. This implies a large amount of
exploring the literature and studying the experts. Without this initial exploration, you can neither know the
range of possibilities of interesting topics nor have a clear idea of what is already known. Most students obtain
their research topics from the loose ends they discover in reading within an area, from an interesting
observation they have made (“I notice that men shut up when a beautiful woman enters the room; I wonder
what the effect of physical attractiveness is on group process?”), or from an applied focus in their lives or
professional work (“I have a difficult time treating these alcoholics and want to discover how best to work with
them”). In short, there is no substitute for immersing oneself in a field of study by having conversations with
leading scholars, advisers, and peers; critically reading the existing literature; and reflecting upon the
implications of professional and personal experiences.
16
Some Guidelines for Topic Selection
Here are some guidelines for deciding whether a topic is appropriate as a dissertation subject.
1. A topic needs to sustain your interest over a long period of time. A study on learning nonsense syllables
under two sets of environmental conditions may sound appealing in its simplicity, but remember Finagle’s first
law of research: If something can go wrong, it will go wrong! Dissertations usually take at least twice as much
time as anticipated, and there are few worse fates than slaving for hour after hour on a project that you abhor.
Remember, too, that all dissertations are recorded and published by the Library of Congress, and you will
always be associated with your particular study.
2. At the other extreme, it is wise to avoid a topic that is overly ambitious and overly challenging. Most
students want to graduate, preferably within a reasonable period of time. Grandiose dissertations have a way
of never being completed, and even the best dissertations end up being compromises among your own
ambition, the wishes of your committee, and practical circumstances. It is not realistic for a dissertation to say
everything there is to say about a particular topic (e.g., the European Union), and you need to temper your
enthusiasm with pragmatism. As one student put it, “There are two types of dissertations: the great ones and
those that are completed!” Sometimes it makes sense to select a research topic on the basis of convenience or
workability and use the luxury of the postgraduate years to pursue more esoteric topics of personal interest.
3. We suggest that you avoid topics that may be linked too closely with emotional issues in your own life. It
always makes sense to choose a topic that is interesting and personally meaningful. Some students, however,
try to use a dissertation to resolve an emotional issue or solve a personal problem. For example, even if you
think you have successfully overcome the personal impact of the death of your child, this is a topic to be
avoided. It will necessarily stir up emotional issues that may get in the way of completing the dissertation.
4. A related issue is selecting a topic in which you have a personal ax to grind. Remember that conducting
research demands ruthless honesty and objectivity. If you initiate a study to demonstrate that men are no
damned good, you will be able neither to allow yourself the sober reflections of good research nor to
acknowledge the possibility that your conclusions may contradict your expectations. It is much better to begin
with a hunch (“I’ve noticed that men don’t do very well with housekeeping; I wonder if that has something to
do with being pampered as children”) and to regard the research as an adventurous exploration to shed light
on this topic rather than as a polemical exercise to substantiate your point of view.
5. Finally, you need to select a topic that has the potential to make an original contribution to the field and
allow you to demonstrate your independent mastery of subject and method. In other words, the topic must be
worth pursuing. At the very least, the study must generate or help validate theoretical understanding in an
area or, in those fields where applied dissertations are permissible, contribute to the development of
professional practice. Some students are put off when they discover that a literature review contains
contradictory or puzzling results or explanations for a phenomenon. However, such contradictions should be
taken not as reasons to steer away from a topic but rather as opportunities to resolve a mystery. When people
17
disagree or when existing explanations seem inadequate, there is often room for a critical study to be
conducted. An opportunity to design a study that resolves theoretical contradictions within a discipline should
not be overlooked.
18
From Interesting Idea to Research Question
Let us assume that you have identified a general area of research and that your choice is based on curiosity and
may involve resolving a problem, explaining a phenomenon, uncovering a process by which something occurs,
demonstrating the truth of a hidden fact, building on or reevaluating other studies, or testing some theory in
your field. To know whether or not the topic is important (significant), you must be familiar with the
literature in the area. In Chapter 4, we present a number of suggestions for conducting a good review and
assessment of the literature. In the meantime, we have noticed that many students have difficulty
transforming an interesting idea into a researchable question, and we have designed a simple exercise to help
in that endeavor.
Researchable questions almost invariably involve a relationship between two or more variables, phenomena,
concepts, or ideas. The nature of that relationship may vary. Research studies generally consist of methods to
explicate the nature of the relationship. Research in the social sciences rarely consists of explicating a single
construct (e.g., “I will look at everything there is to know about the ‘imposter phenomenon’”) or a single
variable (e.g., voting rates in presidential elections).1 Even the presence of two variables is apt to be limiting,
and oftentimes it is only when a third “connecting” variable is invoked that an idea becomes researchable. As a
caveat, however, we acknowledge that research questions that are qualitative rather than quantitative in nature
might not be as focused on the relationship between variables as on “how” processes develop or are
experienced. We will have more to say about this distinction in forthcoming chapters.
An example might help to demonstrate how the introduction of an additional variable can lead to the birth of
a promising study. Let us assume that I am interested in how members of a younger generation perceive the
elderly. At this level, a study would be rather mundane and likely to lead to a “So what?” response. So far, it
implies asking people what they think of the elderly, perhaps using interviews or tests or even behavioral
observations. But we really won’t learn much of value about the nature of perceptions of the elderly in
contemporary society and what influences those perceptions. Introducing a second variable, however, can lead
to a set of questions that have promising theoretical (as well as practical) implications: I wonder what the role
of the media is in shaping social perceptions of the elderly? I wonder if living with a grandparent makes any
difference in how the elderly are viewed? I wonder how specific legislation designed to benefit the elderly has
changed our perception of them? I wonder if there is a relationship between how middle-aged adults deal with
their aging parents and how they view the elderly? The new variables introduced in these potential research
questions are, respectively, the slant of the media, presence or absence of a grandparent, type of legislation,
and treatment of one’s own parents. These variables impart meaning to the research because they offer
suggestions as to what accounts for variability in perceptions of the elderly.
As an example of generating a research question using three primary variables, let’s say that you have inferred
that many women lose interest in sexual relations with their husbands after the birth of a child. At this level,
the proposed study would consist of checking out this hunch by assessing the sexual interest of women
(Variable 1) before and after childbirth (Variable 2). But what would this finding mean? The introduction of a
19
third variable or construct could lead to a much more sophisticated and conceptually meaningful study. An
investigator might ask, “I wonder if the partner’s involvement in parenting makes a difference? What’s the role
of his sexual initiative? How about childbirth complications? Father’s involvement in the birthing? The length
of time they have been married? Presence of other children in the home? Mother’s level of fatigue? Her body
image?” There is no end to the number of interesting questions that can be raised simply by introducing
another variable into the proposed study. This variable would then help to explain the nature of the
relationship between the primary variables. In fact, one could brainstorm a whole list of third variables that
could contribute to a better understanding of the relationship between childbirth and sexuality.
Note that the precise function of the third, or connecting, variable depends on the logic of the conceptual
model or theory underlying the study. In this regard, a distinction can be made between two terms, mediator
and moderator, which play important roles in research questions. A moderator variable pinpoints the
conditions under which an independent variable exerts its effects on a dependent variable. Strictly speaking, a
moderator effect is an interaction effect in which the influence of one variable depends upon the level of
another variable (Frazier, Tix, & Barron, 2004). One commonly employed moderator variable is gender,
which has two levels, male and female. The relationship between provocation and aggression, for example,
may be very different for men and women. The role of context can also be conceptualized as a moderator
variable. The famous Kinsey report on sexual behavior would certainly have generated very different results if
the interviews with participants about their sex lives had taken place in the presence of family members.
Identification of relevant contextual variables has important implications for the design of a study because
such variables will affect the generalizability of research findings.
A mediating variable, on the other hand, tries to describe how or why rather than when or for whom effects will
occur by accounting for the relationship between the independent variable (the predictor) and the dependent
variable (the criterion). The mediator is the mechanism through which the predictor affects the outcome.2 As
such, one can think of mediators as process variables. For example, in the counseling psychology field,
maladaptive perfectionism can be regarded as either a moderating variable or a mediating variable (Wei,
Mallinckrodt, Russell, & Abraham, 2004). Conceptualized as a moderator, attachment anxiety could be seen
to exert negative effects on depressive mood only under conditions of high maladaptive perfectionism (i.e.,
there is a statistical interaction between maladaptive perfectionism and attachment anxiety). Conceptualized
as a mediator, maladaptive perfectionism acts as an intervening variable between attachment anxiety and
depressive mood (i.e., there is an indirect relationship between anxiety and depression). As Wei et al. stated,
It is possible for maladaptive perfectionism to serve as both an intermediate link in the causal chain
leading from attachment insecurity to depressive mood (i.e., as a mediator) and as a variable that alters
the strength of association between attachment insecurity and depressive mood (i.e., as a moderator). (p.
203)
The diagram in Figure 2.1 captures the distinction between moderating and mediating variables in a
theoretical model. In the case of mediation, the mediating variable (maladaptive perfectionism) is placed
20
between attachment anxiety and depressed mood. In the case of moderation, the arrow from maladaptive
perfectionism points to another arrow, that from attachment anxiety to depressed mood, indicating that the
relationship between attachment anxiety and depressed mood depends on the level of maladaptive
perfectionism.
Figure 2.1 The Distinction Between Moderator and Mediator Variables, Represented in a Causal Diagram
Figure 2.2 Classification Plot Representing Moderation of Relationship Between Attachment Anxiety and
Depressed Mood, Moderated by Maladaptive Perfectionism
21
Source: Author created using data from Wei, W., Mallinckrodt, B., Russell, D., & Abraham, W. T.
(2004). Maladaptive perfectionism as a mediator and moderator between adult attachment and
depressive mood. Journal of Counseling Psychology, 51(2), 201–212.
We have illustrated one potential moderated outcome in the classification plot shown in Figure 2.2. Note that
under conditions of low maladaptive perfectionism, there is only a small difference in depression between
those with low and high attachment anxiety; however, under conditions of high maladaptive perfectionism,
those with high levels of attachment anxiety are more likely to be depressed than those with low levels. In
other words, maladaptive perfectionism moderates the relationship between attachment anxiety and
depression, or, put another way, maladaptive perfectionism and attachment anxiety interact.
One research study is not likely to establish and verify all of the important elements of a complex conceptual
model. As one of our colleagues puts it, you would need a video camera to capture the entire Grand Canyon
on film, whereas the dissertation is more like a snapshot, perhaps of a mule and rider descending one small
section of one canyon trail. Yet the proposed model can provide a useful context for current and future
research studies. Most ambitious research studies rely heavily on just such theoretical models.
As you might imagine, a researcher is in no position to test a model of this scope in a single study. For
example, Gerald Patterson and his colleagues (Patterson, DeBaryshe, & Ramsey, 1989) spent many years
developing and testing a model to explain aggressive and deviant behavior among young males. The model
22
hypothesizes that such antisocial behavior can be causally linked to disrupted parental discipline and poor
family management skills. Moreover, the relationship between these two sets of variables is not direct but is
mediated by a network of other variables. The process is thought to begin with parents “training” a child to
behave aggressively by relying on aversive behaviors in both punishment and negative reinforcement
contingencies. The inability of the parents to control coercive exchanges among family members constitutes
“training for fighting,” which leads, in turn, to aggressive behavior and poor peer relationships. This lack of
social skills generalizes to antisocial behavior in the classroom, which makes it next to impossible for the youth
to obtain basic academic skills, thus preparing him poorly to cope with life outside school. Ultimately, this set
of factors leads to high rates of delinquent behavior. An abbreviated summary of one version of the model is
shown in Figure 2.3.
Source: From Patterson, G. R., DeBaryshe, B. D., & Ramsey, E. A developmental perspective on
antisocial behavior. American Psychologist, 44. Copyright © 1989, American Psychological Association.
Reprinted with permission.
Over the years that Patterson and his colleagues (Patterson et al., 1989) spent elaborating the nature of these
relationships, they conducted numerous studies that each constituted a “snapshot” of one aspect of this
complex model, perhaps focusing on a particular set of relationships. For instance, the investigator might ask
whether a relationship exists between physical fighting and poor peer relationships. Each variable would have
to be operationalized, probably by obtaining more than one measure of both fighting and peer relationships.
In Patterson’s work, he asked mothers, peers, and teachers to rate levels of physical fighting because their
perspectives might differ. Likewise, peers, teachers, and self-reports are used to obtain measures of peer
relations. The objective of the study—that is, to determine the nature and form of the relationship between
23
the primary variables—determines the research method that is employed. In the early years of his career,
Patterson focused on the relationships among contextual variables, parental beliefs, parenting practices, and
child outcomes. He concluded that parenting practices—such as discipline, monitoring, problem solving,
involvement, and positive reinforcement—serve as mediating variables between parenting beliefs and attitudes
and the child’s behavior. Once this model was supported by sufficient data, Patterson proceeded to establish
links between children’s behavior problems and subsequent chronic juvenile and adult offending (Reid,
Patterson, & Snyder, 2002).
Whether or not a particular dissertation is designed to test a theory or model derived from the research
literature, we believe that the creation of a visual model, which shows how the network of relevant variables
and constructs may be related to one another, can serve as a powerful tool for guiding the study. Arranging
your ideas spatially helps to organize your thinking, which in turn helps position your proposed study within a
larger framework.
Research models are developed to account for the relationships among variables at a conceptual level and then
used to guide the construction of research designs by which the relationships will be tested, usually, but not
always, using contemporary statistics. The process is iterative so that the models are modified on the basis of
data and then reevaluated in further studies. Two primary types of relationships can be identified and explored
within a causal model (Jaccard & Jacoby, 2010): predictive relationships and causal relationships. A predictive
relationship implies that an association or correlation exists between two (or more) variables without assuming
that one causes the other. For example, we may determine that traveling frequently as a child is related to
(predicts) being more proficient in languages as an adult, without knowing whether travel actually causes
improvement in language skills. Above, we used the terms independent variables and predictor variables more or
less interchangeably. Strictly speaking, however, when the issue is prediction, the relationship is between one
or more predictor variables and a criterion variable (Jaccard & Jacoby, 2010).
Causal relationships imply that one variable “causes” another; that is, changes in the primary variable, usually
referred to as the independent variable, elicit changes in the second variable, the dependent variable or outcome
variable. Although the principle of causation is the foundational bedrock for the experimental method in
social science research, philosophers of science have argued for centuries whether causality can ever be truly
demonstrated. Jaccard and Jacoby (2010) made a persuasive argument that the concept of causality is
ultimately a heuristic that enables us to maintain an organized view of our world and of human behavior. By
inferring causality, we can identify systematic relationships between variables and produce socially significant
changes by manipulating some variables to influence others. Whether or not causality can be definitively
demonstrated, much of contemporary research is conducted to give us confidence in theoretical models that
purport causal relationships.
There are several types of causal relationships, and each type can play a role in developing a causal model.
How causal models can be employed and evaluated in dissertations is a subject for a later chapter (Chapter 6).
How to think about and construct causal models as a way of describing research ideas constitutes much of
Jaccard and Jacoby’s (2010) very useful book, and the following discussion is stimulated by their work. Jaccard
24
and Jacoby observed that most researchers begin by identifying an outcome variable that they want to
understand better. An example might be the level of concern people express for the environment, including by
engaging in behaviors that are environmentally sensitive, such as recycling waste or reducing pollution. The
next challenge is to identify some variables that could potentially influence or relate to environmentally
sensitive behavior. One could imagine a study, for example, that seeks to determine which interventions would
increase the motivation to recycle trash (or, more modestly, just to understand differences between those who
readily recycle and those who do not). Perhaps we predict that having a neighborhood trash collection system
that mandates sorting trash into recyclable and nonrecyclable categories will directly affect environmentally
pro-social behavior. Of course, not all studies begin by identifying a dependent variable; it is also possible to
choose an independent variable and speculate about its effects. For instance, a study might address the
implications of working in a highly polluting industry for health, socioeconomic status, and social
relationships.
Indirect causal relationships have an effect through the influence of an intermediary variable, which we have
referred to above as a mediating variable. Moderated causal relationships are a third type of causal relationship,
again defined earlier in the chapter. Most causal models contain a combination of the various types of
relationships. Models can get very complex because of the number of variables and their subtle relationships
with one another. Thus, simple path diagrams evolve into sophisticated theoretical networks. Many of these
models have been developed over the course of a career by dedicated researchers, such as Gerald Patterson,
who started with studies that explored subsets of variables within a model that then evolved in complexity.
Because computer software allows for the manipulation of multiple variables simultaneously and relatively
effortlessly, researchers usually begin by proposing one or more theoretical models and evaluating them
empirically.
One example of a research model comes from the dissertation of Bill MacNulty (2004), one of our doctoral
students. MacNulty generated this model from existing research literature and then tested it empirically using
a number of well-validated self-report scales. The study employed the schema-polarity model of psychological
functioning to assess how self-schemas (cognitive representations of self and others) influence the experience
of gratitude and forgiveness and whether these variables mediate relationships between self-schemas and
physical health and well-being. The model is summarized in Figure 2.4. The plus and minus signs refer to the
direction of the hypothesized relationships among the variables. Although the results supported most of the
initial hypotheses, the proposed model needed to be amended to accommodate the data. This is typical of the
research enterprise, in which theories and conceptual models are continually tested and refined to serve as
increasingly sophisticated representations of real-life phenomena.
Another dissertation example comes from Jenny Knetig’s (2012) study of active duty military personnel who
are at risk of experiencing symptoms of posttraumatic stress disorder (PTSD). Knetig speculated that being
psychologically minded (a component of what is referred to as having a mentalizing capacity) allows some soldiers
to perceive and interpret cognitive and affective states of themselves and others in a way that might facilitate
resilience and help-seeking behaviors, which, in turn, mitigate how they are affected by severe stress. At the
outset of her study, on the basis of the available literature and her own experience, Knetig postulated that the
25
relationships among these variables might look something like the diagram in Figure 2.5.
Source: From Self-Schemas, Forgiveness, Gratitude, Physical Health, and Subjective Well-Being, by W.
MacNulty, 2004, unpublished doctoral dissertation, Fielding Graduate University, Santa Barbara, CA.
Copyright 2004 by W. MacNulty. Reprinted with permission of the author.
At the conclusion of the study, after collecting her data and performing a number of statistical analyses
(canonical correlation analysis), Knetig amended her proposed model, as shown in Figure 2.6, to reflect the
experiences of her participants more accurately.
In short, the data suggested that soldiers who are more psychologically minded are less apt to conceal their
thoughts and feelings. It also suggested that the relationship between psychological mindedness and
symptoms is mediated by self-concealment.
26
Another random document with
no related content on Scribd:
quills on their upper surface. There are also a few slight lanceolate
dark spots on the sides of the body, and on the tibial feathers.
Length to end of tail 23 1/2 inches, to end of wings 21 1/2, to end of
claws 18 3/4, to carpal joint 5 1/2; extent of wings 51 1/4; breadth of
gape 1 1/8; wing from flexure 17; tail 9 3/4; bill along the ridge 1 5/12;
tarsus 2 4/12; hind toe 1 1/12, its claw 1 5/12; middle toe 2 1/12, its claw
(worn) 11/12. Weight 2 lb., it being much emaciated.
Columba fasciata, Say, in Long’s Exped. to Rocky Mountains, vol. ii. p. 10.
Band-tailed Pigeon, Columba fasciata, Ch. Bonaparte, Amer. Ornith. pl.
viii, fig. 3, vol. i. p. 77.
Columba fasciata, Bonap. Synops. p. 119.
Band-tailed Pigeon, Nuttall, Manual, vol. i. p. 64.
It was omitted to mention that the minute spots on the eggs are
white.
Nuttall’s Dog-wood.
Length to end of tail 13 1/2 inches, to end of wings 11 1/2; wing from
flexure 7 10/12; tail 4 1/2; bill along the ridge 7/12; tarsus 1 2/12; middle
toe 1 1/2/12, its claw 6/12.
Turdus montanus.
PLATE CCCLXIX. Male.
Of this beautiful Thrush, of which a figure not having the black band
running quite across the breast, as is the case in the adult male, is
given by Mr Swainson, in the Fauna Boreali-Americana, Dr
Richardson speaks as follows:—“This species was discovered at
Nootka Sound, in Captain COOK’S third voyage, and male and
female specimens, in the possession of Sir Joseph Banks, were
described by Latham: Pennant has also described and figured the
same male. The specimen represented in this work was procured at
Fort Franklin, lat. 65 1/4°, in the spring of 1826. We did not hear its
song, nor acquire any information respecting its habits, except that it
built its nest in a bush, similar to that of the Merula migratoria. It was
not seen by us on the banks of the Saskatchewan; and, as it has not
appeared in the list of the Birds of the United States, it most probably
does not go far to the eastward of the Rocky Mountains in its
migrations north and south. It may perhaps be more common to the
westward of that ridge.”
Dr Richardson’s conjecture as to the line of march followed by it
has proved to be correct, Dr Townsend and Mr Nuttall having
found it abundant on the western sides of the Rocky Mountains. The
former of these zealous naturalists informs me that he “first found
this Thrush on the Columbia River in the month of October, and that
it becomes more numerous in winter, which it spends in that region,
though some remove farther south. It there associates with the
Common Robin, Turdus migratorius, but possesses a very different
note, it being louder, sharper, and quicker than those of the latter,
and in the spring, before it sets out for its yet unascertained
breeding-place, it warbles very sweetly. It is called Ammeskuk by the
Chinooks.”
Mr Nuttall’s notice respecting it is as follows:—“Of this bird, whose
manners so entirely resemble those of the Common Robin, we know
almost nothing. They probably breed as far north as Nootka, where
they were first seen by the naturalists of Cook’s expedition. On the
Columbia they are only winter birds of passage, arriving about
October, and continuing more or less frequently throughout the
winter. At this time they flit through the forest in small flocks,
frequenting usually low trees, on which they perch in perfect silence,
and are at times very timorous and difficult of approach, having all
the shy sagacity of the Robin, and appearing at all times in a very
desultory manner.”
The numerous specimens of this Thrush in my possession have
enabled me to compare it with Turdus migratorius, and another new
Thrush from Chili. On examining the tail, from the shape of which Mr
Swainson considers this species allied to our Mocking Bird, I found
its form, length, and extent beyond the wings, to correspond almost
exactly with those of the tail of our Robin; and, if it proves true that
the Varied Thrush forms a nest bedded with mud, it will strengthen
my opinion that both these and the Chilian species are as nearly
allied as possible, and therefore ought to be considered as true
Thrushes, of which, to assume the language of systematic writers,
Turdus migratorius is the type in America, whilst Turdus Merula is
that of Europe.
The two figures in my plate were taken from adult males shot in
spring. You will find a figure of the female in Plate CCCCXXXIII.
Turdus nævius, Gmel. Syst. Nat. vol. i. p. 817.—Lath. Ind. Ornith. vol. i. p.
331.
Orpheus meruloides, Thrush-like Mock-bird, Richards. and Swains.
Fauna Bor.-Amer. vol. ii. p. 187.