-
Tending the Wilted Flower: The Role of Psychological Need
Fulfillment in Treatment for Depression
by
Matthew Quitasol
A thesis submitted in conformity with the requirements
for the degree of Masters of Arts
Graduate Department of Psychological Clinical Sciences
University of Toronto
© Copyright by Matthew Quitasol 2016
-
ii
Tending the Wilted Flower: The Role of Psychological Need
Fulfillment
in Treatment for Depression
Matthew Quitasol
Masters of Arts
Graduate Department of Psychological Clinical Science
University of Toronto
2016
Abstract
The present research integrated the principles of
self-determination theory (e.g. Deci & Ryan,
2000) with the cognitive mediation model of depression (e.g.
Whisman, 1993). Participants with
a SCID-IV diagnosis for major depressive disorder were randomly
assigned to 16 weeks of
cognitive therapy or anti-depressant medication. They also
completed indices of depression
severity, neuroticism, and psychological need fulfillment, at
four assessment points (pre-
treatment, week 4, week 8, and week 16). Psychological need
fulfillment increased over the
course of treatment and was negatively correlated with
depression severity. Increases in
psychological need fulfillment predicted decreases in depression
severity over and above the
effects of time, neuroticism, and negative cognitions. The
temporal association between changes
in psychological need fulfilment was bidirectional, and
significantly more pronounced in the
cognitive therapy condition. The association between changes in
psychological need fulfillment
and changes in depression severity was not mediated by reduced
negative cognitions.
-
iii
Acknowledgments
First and foremost, I am grateful to my supervisor, Marc A.
Fournier. His mentorship, guidance,
patience, and insight have been paramount to my development as a
young scientist, as well as my
navigation of the expected and unexpected trials of academia and
clinical graduate work. I am
also grateful to my committee members, R. Michael Bagby and Lena
C. Quilty, whose support
and clinical expertise enabled me grow as a clinical researcher
and push this project beyond my
expectations. I would like to thank my labmates, Stefano I. Di
Domenico, Nic Weststrate, and
Vicki (Mengxi) Dong for vetting my ideas, providing their
invaluable feedback, and their
friendship. I would also like to thank Nina Dhir, Minnie Kim,
Hanan Domloge, and Liz
Pulickeel. Your assistance and advice regarding the
administrative and logistic affairs of
graduate school have helped to lighten the heavy burdens
associated with a clinical program. I
am also grateful to the members of my cohort, Dean Carcone,
Kyrsten Grimes, Le-Anh Dinh-
Williams, and Phil Desormeau. Your friendship and support have
been instrumental in surviving
our forging into clinicians. I would also like to thank my
brothers, Chris Quitasol and Mike
Quitasol. Your constant love and unbreakable support keep me
“humble.” Finally, I am eternally
grateful and indebted to my parents whose unwavering support and
many sacrifices have enabled
me to pursue my passion.
-
iv
Table of Contents
Acknowledgments..........................................................................................................................
iii
Table of Contents
...........................................................................................................................
iv
List of Tables
...................................................................................................................................v
List of Figures
................................................................................
Error! Bookmark not defined.
Chapter 1
Introduction..................................................................................................................1
Chapter 2 Method
..........................................................................................................................9
Chapter 3 Results
.........................................................................................................................13
Chapter 4 Discussion
.....................................................................................................................25
References
.....................................................................................................................................31
Appendix
........................................................................................................................................38
-
v
List of Tables
Table 1. Descriptive statistics and reliabilities.
Table 2. Zero-order correlations between psychological need
fulfillment and depression
collapsing across participants and assessment points.
Table 3. Models of change in psychological need fulfillment and
depression over t ime.
Table 4. Models of Change in Depression as a Function of Changes
in Psychological Need
Fulfillment.
Table 5. Models of Change in Depression as a Function of Changes
in Psychological Need
Fulfillment and Other Personality Variables.
Table 6. Models of Change in Depression as a Function of Changes
in Psychological Need
Fulfillment and Negative Cognitions.
Table 7. Lagged Models of Change in Psychological Need
Fulfillment and Depression Over
Time.
Table 8. Impact of Other Personality Variables on Lagged Models
of Change in Psychological
Need Fulfillment and Depression Over Time.
Table 9. Stepwise Models for Assessing the Mediating Role of
Change in Negative Cognitions
in the Relationship Between Psychological Need Fulfillment and
Depressive Symptoms.
Table 10. The Impact of Treatment Group and Sex on the
Association between Psychological
Need Fulfillment and Depressive Symptoms.
-
1
Chapter 1 Introduction
Carl Rogers (1959) conceptualized psychotherapy as a set of
facilitative conditions that
nurture the client’s inherent tendencies to develop and enhance
their capacity for autonomous
functioning. Just as wilted flowers can be tended back to health
by providing them with sunlight,
nutrient-rich soil, and water, Rogers (1959) maintained that
providing clients with nurturing
conditions, such as unconditional positive regard, was integral
to clients’ psychological growth.
In the present research, we utilized self-determination theory
(SDT), a contemporary framework
which maintains that all living organisms require the
fulfillment of innate psychological needs, to
examine similar nurturing conditions that are theorized to be
integral to psychological growth.
Specifically, we examined the extent to which clients’ basic
psychological needs for autonomy,
competence, and relatedness are fulfilled over the course of
cognitive therapy and
pharmacotherapy for depression, and how the fulfillment of these
needs scaffolds healthy
psychological growth and cognitive change.
Self-Determination Theory
SDT is a macrotheoretical framework of personality, motivation,
and optimal
psychological development (Ryan, 1995; Deci & Ryan, 2000;
Deci & Ryan, 2008). Central to
SDT are three basic psychological needs. Autonomy describes the
experience of one’s behavior
as volitional and reflective of one’s own values and interests;
behavior is experienced as being
self-initiated and self-endorsed (Ryan, 1995; Deci & Ryan,
2000; Deci & Ryan, 2008).
Competence describes the experience of effectance, mastery, and
growth in one’s activities vis-à-
vis the environment (White, 1959; Ryan, 1995; Deci & Ryan,
2000; Deci & Ryan, 2008).
Relatedness describes the experience of feeling cared for and
connected to close others; a sense
of belongingness and the feeling that one matters (Deci &
Ryan, 2000; Deci & Ryan, 2008;
Ryan, 1995).
SDT maintains that healthy psychological functioning is
facilitated by conditions that
support psychological need fulfillment. Just as all plants
flourish when they are provided with
nutrient-rich soil, water, and sunlight, all people thrive and
maximize their psychological
integrity when they experience autonomy, competence, and
relatedness (Deci & Ryan, 2000).
-
Conversely, just as all plants languish and wither when deprived
of water, or when planted in
contaminated soil, so too do all people languish and wither when
their psychological needs are
deprived or frustrated. Indeed, a substantial body of evidence
has demonstrated that fulfillment
of autonomy, competence, and relatedness fosters psychological
growth, integrity, and well-
being; similarly, deprivation or frustration of these
psychological needs degrades psychological
integrity, resulting in psychopathology and ill-being
(Bartholomew, Ntoumanis, Ryan, Bosch, &
Thøgersen-Ntoumani, 2011; Deci & Ryan, 2000; Deci &
Ryan, 2008; Ferrand, Martineent, &
Durmaz, 2014; Reis, Sheldon, Gable, Roscoe, & Ryan, 2000;
Ryan, 1995; Ryan, Deci, Grolnick,
& La Guardia, 2006; Vansteenkiste, Lens, Soenens, &
Luyckx, 2006; Vansteenkiste & Ryan,
2013).
Although SDT has approached the definition of autonomy from the
perspective of self-
governance and the self-authorship of behavior, the operational
definition of autonomy has
historically been heterogeneous across the literature. Notably,
Beck (Beck, Epstein, Harrison, &
Emery, 1983) conceptualized autonomy as a personality construct
reflecting individual
differences in vulnerability to depression. For Beck and
colleagues, the highly autonomous
individual places strong emphasis on personal freedom, mobility,
individuality, achievement, and
is exceptionally sensitive to events perceived as threatening to
these values. Beck maintained that
these qualities make the autonomous individual highly
susceptible to reactive depression. As
such, Beck’s use of the term autonomy is more consistent with
that of an individual difference
characteristic, whereas SDT uses the term autonomy to refer to a
basic psychological need that
characterizes an experiential requirement of all individuals.
Previous research has demonstrated
this heterogeneity in the operational definition of autonomy
(Hmel & Pincus, 2002). In their
psychometric review of various measures of autonomy, Hmel and
Pincus conducted a principle
axis factor analysis and found that autonomy as conceptualized
by SDT, and autonomy as
conceptualized by Beck, were subsumed by separate and distinct
factor structures. These factor
structures were not only separate, but also theoretically
consistent. Indices associated with SDT’s
conceptualization of autonomy as volitional, and reflective of
one’s own values and interests
loaded onto a factor conceptualized as self-governance. This
underlying factor was characterized
by psychological adaptation, self-directness, positive
emotionality, and intrinsic motivation,
qualities that are congruent with the literal meaning of
autonomy (i.e., “self-ruling”). Indices
associated with Beck’s conceptualization of autonomy loaded onto
a factor conceptualized as a
depressogenic vulnerability. In contrast to self-governance,
this underlying factor was
-
characterized by a lack of agentic quality, an inclination
towards negative affect, and
interpersonal detachment tapping a construct reminiscent of a
cognitive personality style that
confers vulnerability to depression (Hmel & Pincus,
2002).
Self-Determination Theory and Treatment for Depression
The role of autonomy is considered by SDT to be an integral
component of motivation
during treatment in psychotherapy (Ryan & Deci, 2008; Ryan,
Lynch, Vansteenkiste, & Deci,
2011). The majority of clinical research applying SDT in the
treatment for depression has
emphasized understanding how clients internalize and freely
endorse treatment-related behaviors
in order to facilitate healthy psychological change. According
to SDT, a depressed client’s
willingness to internalize and participate in treatment-related
behaviors for healthy psychological
change can vary in the degree to which it is autonomously
motivated (Deci & Ryan, 2000). Four
different qualities of internalization lie along this continuum
of autonomous motivation: external
regulation, introjected regulation, identified regulation, and
integrated regulation.
External regulation of treatment-related behaviors is the least
autonomously motivated
form of internalization during therapy. Clients who are
externally regulated participate in their
treatment in order to satisfy an external demand or reward
contingency (Deci & Ryan, 2000).
When demands and reward contingencies regarding participation in
therapy emanate from within
the client, internalization is said to be introjected.
Introjected clients participate in treatment to
avoid feelings of guilt or anxiety (Deci & Ryan, 2000). A
more autonomously motivated form of
internalization during therapy is identified regulation.
Identified clients participate in their
treatment because it is personally important to them and
instrumental to maintaining their mental
health. Finally, the most autonomously motivated form of
internalization is integrated regulation.
Integrated clients fully assimilate and wholeheartedly endorse
their participation in treatment
because healthy psychological functioning is congruent with
their core values and underlying
sense of self. Indeed, previous research has found that
depressed clients report fewer symptoms
and show a higher probability of remission when they are
autonomously motivated to participate
in their treatment (Michalak, Klappheck, & Kosfelder, 2004;
Pelletier, Tuson, & Haddad, 1997;
Ryan & Deci, 2008; Zuroff, Koestner, Moskowitz, McBride,
Marshall, & Bagby, 2007).
According to SDT, a depressed client’s internalization of
treatment-related behaviors
during therapy is actively supported by the clinician through
establishing an autonomy-
supportive treatment context (Ryan & Deci, 2008; Sheldon,
Joiner, Pettit, & Williams, 2003;
-
Ryan, Lynch, Vansteenkiste, & Deci, 2011). Autonomy support
is comprised of three distinct
components (Deci & Ryan, 1985). The first component requires
that the clinician acknowledge
the perspective of the client through validating and honoring
his or her unique world view. The
second component requires that the clinician provide the client
with reasonable and meaningful
choices, such that he or she can freely determine the course of
his or her treatment. Finally, when
choice cannot be provided to the client, it is important for the
clinician to provide the client with
a meaningful rationale for why he or she does not have a choice.
When depressed clients are
provided with the necessary conditions for cultivating
autonomous motivation over the course of
a variety of treatments for depression, they are better able to
engage in treatment-relevant
behaviors and cultivate healthy psychological change (McBride,
Zuroff, Ravitz, Koestner,
Moskowitz, Quilty, & Bagby, 2010; Ryan & Deci, 2008;
Zuroff, Koestner, Moskowitz,
McBride, & Bagby, 2012).
Although autonomy is often discussed with more depth and
elaboration among SDT
researchers because of its long debated controversy (Ryan, Deci,
Grolnick, & La Guardia, 2006),
as well as its importance for describing qualities of motivation
and behavior, SDT researchers
maintain that fulfillment of all three psychological needs is
necessary for optimal motivation and
internalization (Deci & Ryan, 2000; Deci & Ryan, 2008;
Ryan, 1995). Indeed, it is simply not
enough for a behavior or task to be congruent with one’s values,
interests, and goals; one must
also feel capable and competent in the task. Moreover, given
that most behavior does not occur
in an interpersonal vacuum, it is often not enough for one to be
competent and volitional in their
actions; how one acts is also meaningfully connected to other
people, especially close others.
Thus, just as a plant cannot grow by forgoing water for sunlight
or nutrient-rich soil without the
risk of compromising its integrity, SDT maintains that
psychological needs contribute equally to
healthy psychological growth.
Another corollary of the plant metaphor used by SDT to
communicate the importance of
psychological need fulfillment entails that wilted and
languishing plants can be transplanted to
more nurturing conditions in order to cultivate their integrity
and promote healthy growth. By
analogy, therapeutic interventions should provide similar
conditions for individuals with major
depressive disorder to enhance their psychological need
fulfillment in order to promote healthy
and adaptive psychological change (Ryan & Deci, 2008).
However, the majority of research
concerned with psychological need fulfillment in depressed
populations over the course of
treatment has focused largely on the growth manifestations that
emerge as a result of
-
autonomous functioning and autonomy-supportive conditions.
Specifically, clinical research in
SDT has primarily focused on clients’ internalization and
motivations to seek out and maintain
their treatment, and how these motivations affect treatment
outcomes (Lynch, Vansteenkiste, &
Deci, 2011; Ryan & Deci, 2008; Ryan). Moreover, given that
autonomy-supportive conditions
have been found to facilitate the fulfillment of all three
psychological needs (Deci & Ryan, 2000;
Ryan & Deci, 2008), a key limitation of the existing
research bridging SDT with treatment for
depression has been the largely untested assumption that clients
experience increases in
psychological need fulfillment over the course of treatment.
This largely untested assumption—
that psychological needs have either been fulfilled or
frustrated (i.e. a converse error)—is also
present in research bridging SDT with other clinical
populations, including obsessive compulsive
disorder (e.g., Assor & Tal, 2012) and eating disorders
(e.g., Vansteenkiste, Soenens, &
Vandereyecken, 2005). Although some emerging clinical research
using SDT as a framework
has begun to test this assumption (e.g., Verstuyf,
Vansteenkiste, & Soenens, 2012; Verstuyf,
Vansteenkiste, Sonenes, Boone, Mouratidis, 2013), the majority
of SDT research involving
clinical populations has focused less on exploring changes in
psychological need fulfillment or
frustration, and focused more on the outcomes of processes which
imply psychological need
fulfillment or frustration.
Self-Determination Theory and the Cognitive Mediation Model
According to the cognitive mediation model of depression, the
primary process of healthy
psychological change during cognitive therapy comes from change
in negative cognitions (Beck,
Rush, Shaw, & Emery, 1979; Whisman, 1993). Depression is
rooted in a latent depressogenic
self-schema (Beck, 1967; Beck, Rush, Shaw, & Emery, 1979;
Segal & Ingram, 1994) that, upon
activation, causes negative perceptual and cognitive processing
biases. These negative cognitions
can transform banal sad moods into seemingly inescapable
singularities of profound negativity,
resulting in major depressive disorder. Cognitive therapy works
specifically to alter the function,
content, and structure of the depressogenic self-schema.
However, the processes governing the relationships between
therapeutic factors and
healthy psychological change can be very complicated
(Morgenstern & Longabaugh, 2000). Oei
and Free (1995) found that changes in cognitive style were
related to changes in depression
across 44 outcome and process studies of therapy for depression,
but that cognitive change was
not exclusive to cognitive therapeutic interventions. Similarly,
Garratt, Ingram, Rand, and
-
Sawalani (2007) found evidence for cognitive change in the
treatment of depression not only in
cognitive therapy, but also in noncognitive therapy and
psychopharmacology. Given that
cognitive change does not seem to be specific to cognitive
therapy, it is important to identify
additional theoretical frameworks (and their corresponding
processes) with which to expand our
understanding of change processes in treatment. This is
especially important for delineating both
specific and non-specific factors in treatment and how they
contribute to outcomes.
One recent study has demonstrated the promise of applying SDT
principles to the study
of cognitive change in depression. Dwyer, Hornsey, Smith, Oei,
and Dingle (2011) investigated
the role of autonomy over the course of cognitive behavioral
group therapy for depression. They
found that levels of autonomy fulfillment increased following
four weeks of treatment; levels of
autonomy fulfillment were inversely related to depression
severity, an association that was
mediated by reduced negative cognitions. An important limitation
of this research was its
exclusive focus on the need for autonomy to the neglect of the
other two needs. Indeed, given
that cognitive therapy provides opportunities for individuals to
hone personal as well as
interpersonal skills (e.g. Greenberger & Padesky, 1995),
there is no reason to assume that
relatedness fulfillment and competence fulfillment are met any
differently compared to
autonomy fulfillment over the course of treatment.
Overview of the Present Study
In an effort to further clarify the role of SDT in the treatment
of depression and building
on the existing work of Dwyer and colleagues (2011), the present
study will examine how
autonomy, competence, and relatedness fulfillment change over
the course of cognitive therapy
and pharmacotherapy for depression. Moreover, we will examine
the temporal relationship
between changes in psychological need fulfillment and changes in
depression severity. Given
that SDT posits a bidirectional relationship between
psychological need fulfillment and mental
health (Deci & Ryan, 2000), we will examine whether clients
experience an increase in
psychological need fulfilment prior to decreases in depression
severity, subsequent to decreases
in depression severity, or both. Finally, in keeping with the
cognitive meditation model of
depression, we will examine whether changes in negative
cognitions mediate the temporal
relationship between psychological need fulfillment and
depression severity.
Psychological Needs and Personality Change. Insofar as our goal
is to examine the
relationship between changes in psychological need fulfillment
and changes in depression
-
severity, it is also important to demonstrate that changes in
psychological need fulfillment are
not accounted for by other personality variables which have been
found to change over the
course of treatment, such as neuroticism (e.g., Renner Penninx,
Peeters, Cuijpers, Huibers, 2013)
and attachment style (Kinley & Reyno, 2013; Marmarosh &
Tasca, 2013; Maxwell, Tasca,
Ritchie, Balfour, & Bissada, 2014; Travis, Bliwise, Binder,
& Horne-Moyer, 2001). Thus, we
will also examine changes in neuroticism and attachment style as
covariates in the relationship
between changes in psychological need fulfillment and changes in
depression severity.
Psychological Needs and Antidepressant Medication. In addition
to psychotherapy, SDT
has also demonstrated the effect of autonomous motivation on
health outcomes which require
adherence to various medication regiments, including regiments
for antidepressant medication
(ADM; Bruzzese, Idalski Carcone, Lam, Ellis, & Naar-King,
2014; Williams et al, 2009;
Williams, Rodin, Ryan, Grolnick, & Deci, 1998; Zuroff et
al., 2007). However, given the
humanistic foundations of SDT (Deci & Ryan, 2000; Deci &
Ryan, 2008; Ryan, 1995) and our
goal of examining how psychological need fulfillment changes
over the course of treatment, a
more pertinent question regarding the relationship between SDT
and ADM concerns whether or
not ADM is able to produce changes in psychological need
fulfillment that differ from those
produced in psychotherapy. Psychological needs are qualities of
experience that emerge from
interactions with the environment (e.g. social contexts;
Vansteenkiste & Ryan, 2013). It is
possible that ADM can alleviate symptoms that prevent depressed
individuals from actively
engaging with their environment and conditions that facilitate
psychological need fulfillment.
Consequently, we will also exam the role ADM in changes in
psychological need fulfillment.
Research Hypotheses
We propose the following hypotheses:
1. Because psychological need fulfillment promotes healthy
functioning, we predict that
psychological need fulfillment—that is, fulfillment of autonomy,
competence, and relatedness—
will be negatively correlated with depression severity.
2. We predict that psychological need fulfillment will increase
over the course of both cognitive
therapy and pharmacotherapy.
3. We predict that this increase in psychological need
fulfillment will be associated with a
decrease in depressive symptoms over the course of
treatment.
-
4. Given that SDT posits a bidirectional relationship between
psychological need fulfillment and
mental health, we will construct a series of lagged exploratory
models to test both (a) the effect
of psychological need fulfillment on subsequent depression
severity and (b) the effect of
depression severity on subsequent psychological need
fulfillment.
5. We predict that any temporal associations obtained between
change in psychological need
fulfillment and change in depression severity will not be fully
accounted for by change in
attachment style or neuroticism; psychological need fulfillment
will have incremental validity in
its association with depression severity over and above the
effect of change in attachment style
and the effect of change in neuroticism.
6. Insofar as psychological need fulfillment contributes to
existing processes of healthy
psychological change during treatment for depression, we predict
that increases in psychological
need fulfillment will be associated with a decrease in negative
cognitions, and that increases in
psychological need fulfillment will indirectly predict decreases
in depressive symptoms through
reduced negative cognitions.
7. Although the prototypical therapeutic setting is
characterized as a supportive environment
conducive to psychological need fulfillment, antidepressants may
also facilitate depressed
participants’ interaction with need-supportive conditions
thereby enhancing their capacity for
need-fulfillment. However, the frequency with which therapy
clients meet with their therapists
might afford them access to more psychological need fulfillment.
Therefore, we predict that
participants assigned the cognitive therapy will demonstrate
larger increases in psychological
need fulfillment over the course of treatment, as well as a more
robust temporal effect of
psychological need fulfillment on depression severity compared
to participants assigned to
pharmacotherapy.
-
9
Chapter 2 Method
Participants
Sample participants for the present investigation were recruited
through media
advertisements and physician referral by the CAMH Clinical
Research Department for a larger
study examining the cognitive mediation model of depression, and
who met DSM-IV criteria for
major depressive disorder (MDD) based on the Structured
Interview for DSM-IV, Axis I
disorders, Patient version (SCID-I/P; First, Spitzer, Gibbon,
& Williams, 1995). Of the 1,415
potential participants who responded via media advertisements
and physician referrals, 455
individuals expressed interest in participating and consented to
a brief phone interview.
Following the telephone screening phase, 213 individuals were
interested in participation and
eligible for a clinical interview. Of those individuals who
completed a clinic screen, 140 were
eligible for and interested in participation. Participants were
excluded if they met any of the
following criteria: (1) SCID-I/P diagnosis of (a) bipolar
disorders, (b) psychotic disorders, and
(c) substance use disorders; (2) organic brain syndrome; or (3)
current ADM or ECT treatment in
the past 6 months at the time of data collection. A total of 104
patients met criteria for
participation and were assigned to treatment.
Measures
Participants completed six self-report measures.
Balanced Measure of Psychological Needs (BMPN). Psychological
need fulfillment was
indexed using the 18-item BMPN (Sheldon & Hilpert, 2012), a
self-rated measure of perceived
autonomy, competence, and relatedness fulfillment. Participants
were asked to report on their
psychological need fulfillment during the past week. Each
psychological need was indexed via a
6-item subscale consisting of three items that measured need
satisfaction (e.g. “I was successful
completing difficult tasks and projects”) and three items that
measured need dissatisfaction (e.g.
“I did something stupid that made me feel incompetent”).
Participants were asked to indicate the
extent to which they agreed with each of the 18 items using a
7-point scale ranging from 1
(Strongly disagree) to 7 (Strongly agree). The descriptive
statistics of self-rated psychological
need fulfillment are summarized in Table 1.
-
Beck Depression Inventory-II (BDI-II). Participants also
completed the BDI-II (Beck,
Steer, & Brown, 1996) a 21-item self-rated measure of
depression. Participants were presented
with 21 groups of statements; after reading each group of
statements carefully, participants
selected the one statement in each group that best described the
way that they had been feeling
during the past two weeks, including the day on which they were
completing the measure.
Statements for each item (e.g. sadness) were ranked on a 4-point
scale ranging from 0 (e.g. I do
not feel sad) to 3 (e.g. I am so sad or unhappy that I can’t
stand it). The descriptive statistics of
self-rated depressive symptoms are summarized in Table 1.
Cognitive Errors Questionnaire (CEQ). Negative cognitions were
indexed using the
CEQ, a self-rated measure of four types of cognitive errors:
catastrophizing, overgeneralization,
personalization, and selective abstraction (Lefebvre, 1981).
Catastrophizing refers to when the
outcome of an event or an event in and of itself as perceived as
being catastrophic or
unsuccessful. Overgeneralization refers to when the outcome of
one experience or event is
believed to apply to other experiences or events in the future.
Personalization refers to when
personal responsibility is assumed for negative events or
negative events are interpreted to have a
personal meaning. Selective abstraction refers to when attention
is only paid to a single, negative
detail of an event, ignoring the context in which that event
took place.
Participants were presented with 48 different situations that
might occur in daily life (e.g.
being told by your boss that you are being laid off due to a
slowdown in the industry). Each
situation was followed by a possible thought that a person in
that situation might have (e.g. “I
must be doing a lousy job or else he wouldn't have laid me
off"). Participants were asked to
imagine themselves in each of the 48 situations and rate how
similar each possible thought was
to how they would think in that situation using a 5-point scale
ranging from 0 (Not at All) to 4
(Extremely). The descriptive statistics of cognitive errors are
summarized in Table 1.
Dysfunctional Attitudes Scale (DAS). Negative cognitions were
also indexed using the
DAS (Weissman & Beck, 1978), which was originally developed
to test the core premise of
Beck’s (1967) cognitive model of depression: maladaptive
thinking styles and attitudes underlie
vulnerability to depression. Participants were presented with a
list of 40 beliefs and attitudes and
asked to rate the extent to which they agreed that a certain
attitude or belief described how they
thought using a 7-point scale ranging from 1 (Totally Agree) to
7 (Totally Disagree). The
descriptive statistics of self-rated dysfunctional attitudes are
summarized in Table 1.
-
The Revised NEO Personality Inventory (NEO-PI-R). Participants’
personality traits were
assessed using the NEO-PI-R, a questionnaire developed through
rational and factor analytical
methods, to measure the Five Factor Model of Personality:
neuroticism, extraversion,
agreeableness, openness, and conscientiousness (Costa &
McCrae, 1992). Participants were
presented with 240 statements (e.g., “I often feel helpless and
want someone else to solve my
problems” and “I’m a superior person”) and indicated the extent
to which they agreed with each
statement on a 5-point scale. The descriptive statistics of
self-rated personality traits are
summarized in Table 1.
Experiences in Close Relationships-Revised Adult Attachment
Questionnaire (ECR-R).
Participants also completed the ECR-R, a 36-item self-rated
index of attachment style comprised
of two factors, anxiety and avoidance, at Week 0 and Week 16
(Fraley, Waller, & Brennan,
2000). Participants were asked to report on how they generally
felt in their romantic
relationships. Participants indicated the extent to which they
agreed or disagreed with each item
(e.g. “I’m afraid that I will lose my partner’s love”) using a
7-point scale ranging from 1
(Disagree Strongly) to 7 (Agree Strongly). The descriptive
statistics of self-rated attachment style
are summarized in Table 1.
Procedure
Participants were randomly assigned to 16 weeks of cognitive
behavioral therapy or
antidepressant medication (CANMAT, 2001). During data
collection, participants completed the
self-rated Beck Depression Inventory-II (BDI-II; Beck, Steer,
& Brown, 1996) each week.
Participants also completed a battery of measures at four
assessment points (pre-treatment, week
4, week 8, and week 16) to assess severity of depression,
personality traits, and negative
cognitions. Participants also completed a measure of attachment
style during their first and last
assessment point. A measure of psychological need fulfillment
was also included in this battery
later in the data collection. Of the 104 participants who
initiated treatment, 92 completed at least
eight weeks of CBT or pharmacotherapy. Five participants dropped
out of the CBT treatment
group and seven participants dropped out of the pharmacotherapy
treatment group. Of the 92
participants who completed a minimum of eight weeks of
treatment, 51 (30 males, 21 females)
completed measures of psychological need fulfillment. Snijders
and Bosker (2012) have
indicated that Level 2 sample sizes of over 30 participants can
be considered large enough for the
purposes of multilevel modeling, the primary analytic approach
of the present study. Of these 51
-
participants, 29 had been assigned to CBT and 22 had been
assigned to pharmacotherapy. This
sample of participants was 79% Caucasian, and ranged from 18 to
59 years of age (M = 35.30,
SD = 9.83). Participants’ average level of education was
moderate (total years of education, M =
16.18, SD = 1.90). Not all participants indicated their total
annual household income; however,
75% of participants indicated that their annual household income
was more than $20,000 per
year. Participants who completed measures of psychological need
fulfillment reported fewer
dysfunctional attitudes at intake than those participants who
dropped out of treatment or did not
complete treatment, t (39.95) = -2.4065, p < .05. In
comparison to those who completed
measures of psychological need fulfillment, participants who did
not complete measures of
psychological need fulfillment reported at intake significantly
more depressive symptoms, t
(79.56) = -2.09, p < .05, made more personalization cognitive
errors, t (81.65) = -2.39, p < .05,
and were significantly more conscientious, t (62.63) = -2.8529,
p < .01.
-
Chapter 3
Results
Testing Hypothesis 1: Is Psychological Need Fulfillment
Significantly Negatively
Correlated with Depression?
Table 1 contains the descriptive statistics for the BMPN,
BDI-II, DAS, CEQ, and NEO-
PI-R for all four assessment points, as well as the descriptive
statistics for the ECR-R for the first
and last assessment points. Table 2 contains the zero-order
correlations for these measures
collapsed across assessment points and participants. In
accordance with our first hypothesis,
psychological need fulfillment was significantly and negatively
correlated with self-reported
depressive symptoms (Autonomy, r = -.27, p < .05; Competence,
r = -.86, p < .01, Relatedness, r
= -.73, p < .01, Composite Psychological Need fulfillment, r
= -.78, p < .01). Psychological need
fulfillment was also differentially related to cognitive
distortions. Autonomy was negatively
correlated with overgeneralization (r = -.29, p < .05).
Competence was negatively correlated with
dysfunctional attitudes (r = -.72, p < .01), selective
abstraction (r = -.61, p < .01), and
personalization (r = -.32, p < .05). Relatedness was
significantly correlated with dysfunctional
attitudes (r = -.56, p < .01) and selective abstraction (r =
-.46, p < .01). Composite psychological
need fulfillment was significantly correlated with dysfunctional
attitudes (r = -.53, p < .01),
selective abstraction (r = -.49, p < .01), and
overgeneralization (r = -.24, p < .05). In accordance
with the cognitive model of depression (Beck, 1967; Weissman
& Beck, 1978), self-reported
depression was significantly correlated with dysfunctional
attitudes (r = .56, p < .01) as well as
for the selective abstraction scale from the CEQ (r = .24, p
< .05). Contrary to the cognitive
model of depression, no other measure of negative cognitions was
significantly correlated with
self-reported depression.
Finally, there were a number of significant intercorrelations
among the Big Five
personality traits. Although Costa and McCrae (1992b) have
argued that correlations among the
Big Five are method artifacts, the Big Five in the present study
were intercorrelated in patterns
that were consistent with previous findings (e.g., John &
Srivastava, 1999; Krueger, Caspi,
Moffit, Silva, & McGee, 1996; Yik & Russell, 2001).
-
Testing Hypothesis 2: Does Psychological Need Fulfillment
Increase Over the Course of
Treatment for Depression?
To assess the within-person change in psychological need
fulfillment and depression,
conventional growth curve analyses were conducted to test each
of the primary hypotheses in the
present study. Growth curve analysis is a technique that
utilizes hierarchical linear models (Bryk
& Raudenbush, 1987; Raudenbush, Bryk, Cheong, Congdon, &
du Toit, 2004; Singer & Willett,
2003; Snijders & Bosker, 2012; Tasca & Gallop, 2009).
Growth curve analysis offers a multitude
of advantages over other analytic techniques (Byrne &
Crombie, 2003; Singer & Willet, 2003;
Tasca & Gallop, 2009). Notably, growth curve analysis can
accommodate unsystematic missing
data, participants with differing numbers of assessment points,
and uneven spacing in the data
collection schedule. Given that such instances of “unbalanced
data” (Singer & Willett, 2003, p.
146) are common in ambulatory, outpatient treatment settings,
such as the one used in the
present study, growth curve analysis was an integral technique
for assessing change in
psychological need fulfillment over the course of treatment for
depression. The analyses for the
present study were conducted in R version 3.3.0 (R Development
Core Team, 2016) using the
nlme package (Pinheiro, Bates, DebRoy, Sarkar, & R
Development Core Team, 2010).
Prior to constructing the proposed growth curve models, we
assessed the extent to which
participants’ data were appropriate for conducting multilevel
analysis. We calculated the
intraclass correlation coefficients (ICC) for each self-report
measure completed by participants in
order to determine the degree of non-independence for
participants’ self-reports across all four
assessment points (Hayes, 2006). The ICCs for each variable
indexed in the present study are
presented in Table 1. The ICCs presented in Table 1 suggest that
there are important
dependencies in the data that need to be accounted for, but that
these dependencies are not so
great as to suggest that there is no session-to-session
variation to study.
Growth curve models were constructed to examine changes in
psychological need
fulfillment across all four assessment points. The results of
these models are presented in Table
3. Model 1 represents change in the experience of autonomy over
time, Model 2 represents
change in the experience of competence over time, Model 3
represents change in the experience
of relatedness over time, and Model 4 represents change in the
experience of overall
psychological need fulfillment over time. The coefficients for
Models 1 through 3 suggest that
the typical participant experienced increased autonomy
fulfillment, b = 0.04, SE = 0.01, t (109) =
3.48, p < .001, competence fulfillment, b = 0.07, SE = 0.013,
t (108) = 5.03, p < .001, and
-
relatedness fulfillment, b = 0.04, SE = 0.01, t (109) = 3.04, p
< .005, across assessment points as
he or she moved through treatment for depression. Similarly, the
coefficients for Model 4
suggests that psychological need fulfillment as a whole
increased across assessment points, b =
0.05, SE = 0.01, t (108) = 4.86, p < .001. Inspection of the
95% confidence intervals for Models
1 through 4 suggest that autonomy, competence, relatedness, and
overall psychological need
fulfillment did not significantly differ from one another in
their rates of change across
assessment point.
Testing Hypothesis 3: Do Changes in Psychological Need
Fulfillment Relate to Changes in
Depression Across Assessment Points?
Having established that psychological need fulfillment increased
over the course of
treatment, we were able to test whether or not changes in
psychological need fulfillment were
related to changes in depression severity. Before we could test
this hypothesis, it was important
to examine whether or not severity of depression decreased over
the course of treatment. Model
5 in Table 3 suggests that depressive symptoms improved over the
course of treatment, b = -1.06,
SE = 0.09, t (456) = -12.00, p < .001. Because severity of
depression significantly decreased over
the course of treatment, we were able to test whether or not
changes in depression severity were
associated with changes in psychological need fulfillment.
Models 6 through 9 reflect the
associations between changes in psychological need fulfillment
and changes in depression
severity. Autonomy fulfillment, b = -5.86, SE = 1.24, t (107) =
-4.72, p < .001, competence
fulfillment, b = -6.97, SE = 0.89, t (106) = -7.80, p < .001,
relatedness fulfillment, b =-5.81, SE
=1.19, t (107) = -4.87, p < .001, and overall psychological
need fulfillment, b = -9.43, SE = 1.23,
t(106) = -7.65, p < .001, were all significantly and
negatively associated with depressive
symptoms. Inspection of the 95% confidence intervals for Models
6 through 9 revealed that the
slopes for autonomy, competence, relatedness, and composite
psychological need fulfillment
were not significantly different from one another in terms of
predicting change in depressive
symptoms. Because all three psychological needs increased over
the course and treatment and
were significantly related to decreased depressive symptoms
across assessment points, the
remainder of our analyses were conducted exclusively using the
composite for psychological
need fulfillment.
In testing the relationship between psychological need
fulfillment and depression severity
over the course of psychotherapy and psychopharmacology, it was
essential to account for the
-
passage of time in our growth curve models. Indeed, the popular
adage that time can heal all
wounds is reflected in the ubiquitous issue of regression to the
mean, a complicating factor in the
assessment of change (Kazdin, 2007, Lambert and Ogle, 2004). We
wanted to account for this
natural change process in our analyses. The results of Model 10
represent the relationship
between psychological need fulfillment and depression severity
over the course of our four
assessment points controlling for the impact of the passage of
time in the form of number of
weeks since the first assessment point. According to this model,
the passage of time was
significantly associated with a decrease in depressive symptoms,
b = -0.79, SE = 0.09, t (105) = -
8.58, p < .001. However, after accounting for this effect of
time, psychological need fulfillment
remained significantly negatively associated with depression
severity over the course of
treatment, b = -6.40, SE = 1.12, t (105) = -5.69, p <
.001.
Beyond regression to the mean, we also wanted to demonstrate the
incremental validity
of psychological need fulfillment in the prediction of
depression severity over and above the
additional changes known to occur over the course of
psychotherapy and pharmacotherapy.
Specifically, previous research has shown that personality
change can accompany symptom
change over the course of treatment (e.g., Renner Penninx,
Peeters, Cuijpers, & Huibers, 2013).
Notably, both attachment style and trait neuroticism have been
implicated not only in the
vulnerability to depression but have been shown to change over
time (Kinley & Reyno, 2013;
Marmarosh & Tasca, 2013; Maxwell, Tasca, Ritchie, Balfour,
& Bissada, 2014; Travis, Bliwise,
Binder, & Horne-Moyer, 2001). We therefore wanted to
demonstrate that the observed
association between changes in psychological need fulfillment
and changes in depression
severity were not better accounted for by change in these
personality variables. In order to test
this, we developed an additional set of growth curve models.
Table 5 summarizes the results of
these models. In order to account for the role of change in
attachment and neuroticism in
depression severity, we first needed to test whether attachment
and neuroticism changed
significantly across assessment points. Models 11, 12, and 13 in
Table 3 represent participants’
changes in neuroticism, attachment anxiety, and attachment
avoidance over the course of
assessment. The results of these model indicate that only
neuroticism, b = -1.00, SE = 0.28, t (67)
= -3.62, p < .001, and attachment anxiety, b = -0.03, SE =
0.01, t (41) = -3.35, p < .01, showed a
significant decrease over the course of treatment. Given the
results of Models 11 and 12, we then
constructed a series of models to test if changes in neuroticism
and changes in attachment
anxiety were significantly associated with changes in depression
severity. Model 14 represents
-
the association between neuroticism and depression severity and
Model 15 represents the
association between attachment anxiety and depression severity.
Neuroticism was found to be
significantly associated with depression severity over and above
the impact of the passage of
time, b = 0.33, SE = 0.06, t (66) = 5.27, p < .01. Similarly,
attachment anxiety, b = 3.904871, SE
= 1.51, t (38) = 2.59, p < .05, was significantly associated
with depression severity over and
above the impact of the passage of time.
Having established that both attachment and neuroticism were
significantly associated
with depression, two additional models were constructed to test
whether psychological need
fulfillment was associated with depression severity over and
above attachment and neuroticism.
Model 16 in Table 5 represents the association between
psychological need fulfillment and
depression accounting for both the passage of time and
neuroticism. The results of this model
suggest that psychological need fulfillment was significantly
negatively associated with
depression severity across assessment points over and above the
effect of time and changes in
neuroticism, b = -4.826842, SE = 1.2730005, t (42) = -3.79, p
< .001. Model 17 in Table 5
represents the relationship between psychological need
fulfillment and depression accounting for
both the passage of time and attachment anxiety. The results of
this model suggest that
psychological need fulfillment was not significantly associated
with depression severity over the
course of treatment after accounting for changes in attachment
anxiety and the effect of time, b =
-2.67, SE = 2.054, t (19) = -1.30, p = 0.21.
Given that cognitive change has been shown to be a significant
component of treatment
for depression (Garratt, Ingram, Rand, & Sawalani, 2007) we
wanted to demonstrate that
psychological need fulfillment could contribute to changes in
depressive symptoms above and
beyond this existing process of change. With this mind, and
building off of the correlations
between the CEQ selective abstraction scale and depressive
symptoms and the correlations
between the DAS and depressive symptoms, we constructed a series
models examining the
relative contributions of cognitive change and psychological
need fulfillment to changes in
depressive symptoms. Table 6 displays the results of these
models. First, we wanted to test
whether or not negative cognitions changed over the course of
treatment. The coefficients in
Models 18 and 19 indicate that both dysfunctional attitudes, b =
0.03, SE = 0.01, t (131) = -4.53,
p < .001, and selective abstraction, b = -0.17, SE =
0.0337979, t (131) = -4.89, p < .001)
decreased significantly over the course of treatment. Models 20
and 21 represent that these
decreases in participants’ dysfunctional attitudes, b = 6.60, SE
= 1.41, t (128) = 4.67, p < .001,
-
and cognitive errors, b = 0.93, SE = 0.29, t (128) = 3.25, p
< .01, were significantly associated
with decreases in their depressive symptoms over and above the
effect of time. Having
demonstrated that decreases in dysfunctional attitudes and
decreases in cognitive errors were
associated with decreases in depressive symptoms, we wanted to
test whether increases in
psychological need fulfillment contributed to changes in
depressive symptoms above and beyond
that of changes in negative cognitions. Models 22 and 23 in
Table 6 were constructed to test this.
The results of these models indicated that increases in
psychological need fulfillment contributed
to decreases in depressive symptoms over and above the effects
of decreases in dysfunctional
attitudes, b = -5.31, SE = 1.03, t (89) = -5.17, p < .001,
and selective abstraction, b = -5.33, SE =
0.98, t (90) = -5.44, p < .001.
Testing Hypothesis 4: Is There a Significant Temporal
Relationship Between Changes in
Psychological Need Fulfillment and Changes in Depression
Severity?
In order to examine the temporal relationship between changes in
psychological need
fulfillment and changes in depression severity over the course
of CBT and pharmacotherapy for
depression, a series of models were constructed using lagged
analyses (Singer and Willett, 2003).
Lagged modeling makes it possible to grapple with issues
surrounding state dependence and
reciprocal causation. One concern in the present study was the
possibility that participants’
psychological need fulfillment may have been state dependent on
the severity of their depressive
symptoms at each assessment point. Alternatively, it is also
possible that participants’ depressive
symptoms may have been state dependent on their psychological
need fulfillment at each
assessment point. The advantage of lagged modeling is that it
allows for testing and confirming
the direction of associations. Although lagged models cannot
determine the causal pathways in
longitudinal relationships, a conclusion reserved for the
addition of experimental controls,
lagging predictor variables does allow one to establish the
temporal precedence of an association.
According to Singer and Willett (2003), examining the temporal
precedence of an association
requires lagged models to be constructed based on theory. SDT
maintains that psychological
need fulfillment is integral to psychological health and growth.
As noted previously, the existing
literature surveying the relationship between psychological need
fulfillment, internalization, and
psychopathology suggests that environments which facilitate
psychological need fulfillment are
what lead to psychological health and growth (Deci & Ryan,
2000; Deci & Ryan, 2008; Ryan,
1995; Ryan, Deci, & Vansteenkiste, 2016). Given the
assumption that psychological need
-
fulfillment is a prerequisite for health, it is possible that
participants in the present study
experienced changes in their psychological need fulfillment
prior to experiencing changes in
their depressive symptoms. Alternatively, SDT maintains that
psychological need fulfillment
does not abide by a drive reduction model typically associated
with conceptualizations of needs
(Hull, 1943). Specifically, individuals are not impelled to
experience autonomy, competence, and
relatedness in order to satisfy or reduce a drive in the same
way they are impelled to eat in order
to reduce hunger. Rather, psychological need fulfillment is
drive-inducing (Deci & Ryan, 2000;
Ryan, Deci, & Vansteenkiste, 2016), where the experiences of
autonomy, competence, and
relatedness energize motivation and broaden individuals’
capacities to seek out additional
psychological need fulfillment. In this vein, it is also
possible that participants in the present
study experienced changes in their depressive symptoms prior to
experiencing changes in
psychological need fulfillment.
Given that SDT proposes a bidirectional relationship between
psychological need
fulfillment and psychological health, we constructed two lagged
models to test this. The results
of these models are presented in Table 7. Model 24 in Table 6
represents the association between
changes in psychological need fulfillment and subsequent changes
in depressive symptoms at a
lag of one assessment point. The results of this model indicate
that there was a significant
negative association between changes in psychological need
fulfillment and subsequent changes
in depressive symptoms, b = -4.32, SE = 1.63, t(64) = -2.65, p
< .05. Model 25 in Table 7
represents the converse association: The association between
changes in depressive symptoms
and subsequent changes in psychological need fulfillment. The
result of this model suggest that
there was a significant negative relationship between changes in
depressive symptoms and
subsequent changes in psychological need fulfillment, b = -0.03,
SE = 0.01, t (79) = -2.73, p <
.01. Collectively, these two models indicate that participants’
depressive symptoms not only
decreased prior to associated increases in their psychological
need fulfillment, but also that
participants’ experience of psychological need fulfillment
increased prior to associated decreases
in their depressive symptoms. Models 26 and 27 in Table 7 were
constructed to test whether
these temporal relationships between psychological need
fulfillment and depression severity
remained significant after accounting for the effect of time.
The results of these models indicated
that neither of these lagged associations were significant over
and above the passage of time.
-
Testing Hypothesis 5: Do Other Personality Variables Impact the
Temporal Relationships
Between Changes in Psychological Need Fulfillment and Changes in
Depression Severity?
Beyond exploration of the temporal relationship between changes
in psychological need
fulfillment and changes in depressive symptoms, we wanted to
explore how both neuroticism
and attachment contribute to this relationship. In this vein we
constructed an additional set of
models using neuroticism and attachment anxiety as covariates in
the temporal association
between psychological need fulfillment and depression severity.
First, given the bidirectional
temporal relationship between psychological need fulfillment and
depression severity, it was
necessary to examine psychological need fulfillment and its
respective relationships with
neuroticism and attachment anxiety. Models 28 and 29 in Table 8
were constructed to test these
respective relationships. The results of these models indicated
that participants’ changes
neuroticism, b = -0.03, SE = 0.008, t (43) = -3.55, p < .01,
and changes in attachment anxiety, b
= -0.57, SE = 0.16, t (21) = -3.68, p < .01) were both
significantly and negatively associated with
their change in psychological need fulfillment across assessment
points over and above the effect
of time.
Given the findings suggested by these models, it was possible to
explore the relative
contributions of both changes in psychological need fulfillment
to subsequent changes in
depressive symptoms, and changes in depressive symptoms to
subsequent changes in
psychological need fulfillment above and beyond that of
neuroticism and attachment anxiety.
Models 30 and 31 were constructed to test the association
between participants’ increases in
psychological need fulfilment and subsequent decreases in
depressive symptoms, accounting for
their decreases in both neuroticism and attachment anxiety over
the course treatment,
respectively. The results of both of these models indicate that
the association between increases
in psychological need fulfillment and subsequent decreases in
depressive symptoms was no
longer significant after statistically controlling for changes
in neuroticism, b = 1.03, SE = 2.90, t
(9) = 0.36, p = 0.73, and attachment anxiety, b = -0.22, SE =
6.13, t (13) = -0.04, p = 0.97.
Models 32 and 33 test the reverse of this association: The
relationship between participants’
decreases in depressive symptoms and subsequent increases in
psychological need fulfillment,
after accounting statistically for the effects of neuroticism
and attachment anxiety. The
coefficients in both of these models demonstrate that the
association between decreases in
depressive symptoms and subsequent increases in psychological
need fulfillment was no longer
significant after accounting statistically for the effects of
neuroticism, b = 0.003, SE = 0.017,
-
t(28) = 0.16, p = 0.8777, and attachment anxiety, b = -0.05, SE
= 0.03, t(13) = -1.76, p = .48).
Together, these four models indicate that the bidirectional
temporal relationship between changes
in psychological need fulfillment and changes in depressive
symptoms were not significant over
and above change in personality variable over the course of
treatment.
Testing Hypothesis 6: Is the Relationships Between Increases in
Psychological Need
Fulfillment and Decreases in Depression Severity Mediated by
Change in Negative
Cognitions?
We hypothesized that increases in psychological need fulfillment
would be associated
with a decrease in negative cognitions, and that increases in
psychological need fulfillment
would indirectly predict decreases in depressive symptoms
through reduced negative cognitions.
Given the zero-order correlations presented in Table 2, we chose
to focus on the DAS as our
primary index of negative cognitions. In order to test this
hypothesis, we implemented a variant
of the causal steps approach (Baron & Kenny, 1986) amended
for 1-1-1 multilevel mediation
(Zhang, Zyphur, & Preacher, 2009), because all of the
predictor (i.e., psychological need
fulfillment), mediator (i.e., negative cognitions), and outcome
variables (i.e., depression severity)
varied across all participants. Because this type of mediation
model confounds Level 1 variables
with Level 2 variables, we included both the aggregated and
within-person-mean-centered values
for all relevant predictors (Zhang, Zyphur, & Preacher,
2009). All multilevel models used an
unstructured covariance matrix and the between-within method of
estimating degrees of
freedom.
Three multilevel models were constructed. First, participants’
negative cognitions were
modeled as a function of aggregated psychological need
fulfillment and within-person-mean-
centered psychological need fulfillment, with a random slope for
the influence of within-person-
mean-centered psychological need fulfillment estimated for each
participant. The results of this
model are represented by Model 34 in Table 9. This model
revealed a significant negative effect
of aggregated psychological need fulfillment on participants’
negative cognitions, b = -0.50, SE
= 0.12, t (47) = -4.29, p < .001), and a significant effect
of within-person-mean-centered
psychological need fulfillment on participants’ negative
cognitions, b = -0.42, SE = 0.09, t (93) =
-4.94, p < .001. At the lowest level, this first model
reduced prediction error by a very large
amount, R21 = .70; at the second level, this model also reduced
prediction error by a large
amount, R22 = .77. Second, change in depressive symptoms were
modelled as a function of
-
aggregated psychological need fulfillment and
within-person-mean-centered psychological need
fulfillment, with a random intercept for each participant. Model
35 in Table 9 depicts the results
of this model. This model revealed a significant negative effect
of aggregated psychological need
fulfillment on depressive symptoms, b = -9.01, SE = 1.13, t (49)
= -7.95, p < .001, and a
significant effect of within-person-mean-centered psychological
need fulfillment on depressive
symptoms, b = -9.33, SE = 1.23, t (106) = -7.61, p < .001. At
the lowest level, the second model
reduced prediction error by a large amount, R21 = .47; at the
second level, this model also
reduced prediction error by a large amount, R22 = .57. Finally,
depressive symptoms were
modeled as a function of aggregated psychological need
fulfillment, within-person-mean-
centered psychological need fulfillment, aggregated negative
cognitions, and within-person-
mean-centered negative cognitions, with a random slope for the
influence of within-person-
mean-centered negative cognitions for each participant. The
final model, represented by Model
36 in Table 9, revealed a significant effect of aggregated
psychological need fulfillment on
depressive symptoms, b = -7.69, SE = 1.22, t (46) = -6.32, p
< .001, a significant effect of
within-person-mean-centered psychological need fulfillment on
depressive symptoms, b = -6.61,
SE = 1.11, t (90) = -5.97, p < .001, a nonsignificant effect
of aggregated negative cognitions on
depressive symptoms, b = 1.79, SE = 1.25, t (46) = 1.43, p =
0.16, and a significant effect of
within-person-mean-centered negative cognitions on depressive
symptoms, b = 6.73, SE = 1.65, t
(90) = 4.07, p < .001. At the lowest level, the final model
reduced prediction error by a large
amount, R21 = .54; at the second level, the final model also
reduced prediction error by a large
amount, R22 = .57. According to Zhang, Zyphur, and Preacher
(2009), the criteria for testing
mediation was observed at Level 1 because the effect of
within-person-mean-centered
psychological need fulfillment was significant in the first
model, b = -0.42, SE = 0.09, t (93) = -
4.94, p < .001, and the effects of
within-person-mean-centered psychological need fulfillment, b
= -6.61, SE = 1.11, t (90) = -5.97, p < .001, and
within-person-mean-centered negative
cognitions, b = 6.73, SE = 1.65, t (90) = 4.07, p < .001,
were significant in the final model.
To test for mediation, first we had to ascertain how consistent
our mediation model was
across the Level 2 groups and so we computed the population
covariance of the random slopes of
the indirect path from psychological need fulfillment to
negative cognitions, σab = -0.018. A
significance test of the correlation between the random slopes
indicated that the population
covariance did not differ from zero, rab = -0.04, p = 0.78,
which implies that the indirect effect of
psychological need fulfillment on depressive symptoms was
consistent for all participants. The
-
population covariance was used to conduct a Sobel Test with the
Aroian correction, ab = -2.83,
Sobel z = -1.71, p = 0.09, suggesting that the indirect effect
was not significantly different from
zero. Thus, change in negative cognitions was not found to
mediate the association between
changes in psychological need fulfillment and changes in
depression severity. The complete
results our proposed mediation model are depicted in Figure
1.
Testing Hypothesis 7: Does the Temporal Relationships Between
Changes in Psychological
Need Fulfillment and Changes in Depression Severity Differ
Across Treatment and Client
Variables?
Participants’ random assignment to pharmacotherapy and cognitive
therapy provided us
with the opportunity to compare the temporal relationship
between changes in psychological
need fulfillment and depression severity across treatment
conditions. To test for the effect of
treatment group, participants who were randomly assigned to
pharmacotherapy were dummy
coded with 1 and participants who were random assigned to
cognitive therapy were dummy
coded with a 0.
We first tested two models to examine if treatment group
moderated overall change in
psychological need fulfillment and depressive symptoms over the
course of treatment. Models 37
and 38 depict the results of these tests. According to the
coefficients of these models, treatment
group did not moderate overall change in psychological need
fulfillment, b = -0.03, SE = 0.02, t
(107) = -1.33, p = 0.18, nor did it moderate overall change in
depressive symptoms, b = 0.21, SE
= 0.13, t (455) = 1.52, p = 0.13.
Models 39 and 40 examine the effects of treatment group on the
association between
changes in psychological need fulfillment and subsequent changes
in depressive symptoms, and
the effects of treatment group on the association between
changes in depressive symptom and
subsequent changes in psychological need fulfillment. In
accordance with our hypothesis, there
was a significant interaction effect for treatment group in both
lagged models. Specifically,
participants receiving cognitive therapy demonstrated stronger
associations between increases in
psychological need fulfillment and subsequent changes in
depressive symptoms, b = 7.27, SE =
2.58, t (62) = 2.82, p < .01), as well as stronger
associations between decreases in depressive
symptoms and subsequent increases in psychological need
fulfillment, b = 0.05, SE = 0.02, t (77)
= 2.88, p < .01. Examination of the simple slopes for Model
39 revealed a significant effect for
lagged psychological need fulfillment on subsequent depressive
symptoms in the cognitive
-
therapy condition, b = -5.90, SE = 1.34, t (48) = -4.41, p <
.01, and a nonsignificant effect for
lagged psychological need fulfillment on subsequent decreases in
depressive symptoms in the
antidepressant medication condition, b = 1.37, SE =1.76, t (21)
= 0.78, p = 0.44. Similarly,
examination of the simple slopes for Model 40 revealed a
significant effect for lagged depressive
symptoms on subsequent psychological need fulfillment in the
cognitive therapy condition, b = -
0.04, SE = 0.01, t (48) = -2.78, p < .01, and a
nonsignificant effect for lagged depressive
symptoms on psychological need fulfillment in the antidepressant
medication condition b = .01,
SE = -1.43, t (48) = -0.01, p = 0.50. Indeed, the reciprocal
lagged relationship between changes
in psychological need fulfillment and changes in depressive
symptoms was more robust for
participants in the cognitive therapy condition than for
participants in the antidepressant
medication condition.
In addition to testing differences across treatment groups, we
also constructed models to
test for the effects of client variables such as sex, number of
years of education, and income on
changes in psychological need fulfillment. The results of these
models indicated that increases in
psychological need fulfillment over the course of treatment did
not differ with respect to sex, b =
-0.004, SE = 0.02, t (107) = -0.23, p = 0.81; number of years of
education, b = -0.004, SE = 0.01,
t (107) = -0.78, p = 0.44; or income, b = 0.04, SE = 0.02, t
(107) = 1.60, p = 0.11. These findings
indicate that increases in psychological need fulfillment over
the course of treatment were
consistent across clients of varying backgrounds.
-
Chapter 4
Discussion
The present study sought to examine the relationship between
changes in psychological need
fulfillment and changes in depressive symptoms over the course
of cognitive therapy and ADM.
Our findings indicated that over the course of treatment,
participants showed increases in
autonomy, competence, and relatedness, and that these increases
in need fulfillment were
commensurate across clients with varying backgrounds. This
increase in participants’
psychological need fulfilment was significantly and negatively
associated with depressive
symptoms over the course of treatment. This relationship was
also significant above and beyond
the passage of time as well as changes in neuroticism over the
course of treatment; changes in
dysfunctional attitudes; changes in cognitive errors; but not
changes in attachment. Moreover,
lagged growth curve models indicated that the relationship
between changes in psychological
need fulfillment and changes in depressive symptoms over time is
bidirectional. Over the course
of psychotherapy and ADM, not only did increases in
psychological need fulfillment predict
subsequent decreases in depressive symptoms, but also decreases
in depressive symptoms
predicted subsequent increases in psychological need
fulfillment. However, this bidirectional
relationship was no longer significant after accounting for the
passage of time, changes in
neuroticism, or changes in attachment. This bidirectional
relationship did not differ between
males or females. However, it was more pronounced in
participants who had been randomly
assigned to ADM. Given the inconclusive specificity of cognitive
change in treatments for
depression, we had hypothesized that the observed relationship
between increases in
psychological need fulfillment and decreases in depressive
symptoms would be mediated by a
reduction in negative cognitions. Contrary to our hypothesis,
however, the indirect effect of
psychological need fulfillment on changes in depressive symptoms
via change in negative
cognitions was nonsignificant.
Changes in Psychological Need Fulfillment Over Time
In the present study, participants experienced increases in
autonomy, competence, and
relatedness over the course of four assessment points during
treatment for depression. This
finding was predicted based on previous research that has
identified changes in patients’
-
experiences of autonomy over the course of four weeks of
cognitive behavioral therapy (Dwyer,
Hornsey, Smith, Oei, & Dingle, 2011). The present study
extends these findings, demonstrating
that not only does autonomy increase over the course of
treatment for depression, but so does
competence and relatedness. Moreover, the present study extends
these changes in psychological
need fulfillment further by demonstrating that autonomy,
competence, and relatedness increase
over the course of both cognitive therapy and ADM. The finding
that psychological need
fulfillment increases over the course of treatment is also
meaningful because it confirms much
about what has largely been assumed in the literature bridging
the treatment of psychopathology
and SDT. As previously mentioned, much of the work bridging
these two areas of research
places emphasis on qualities of motivation and internalization
and how they impact treatment
outcomes, presupposing the fulfillment or frustration of
psychological needs. The present study
therefore adds to a growing body of research which documents the
role of changes in
psychological need fulfillment in understanding psychopathology
and its treatment (Dwyer,
Hornsey, Smith, Oei, & Dingle, 2011; Verstuyf,
Vansteenkiste, & Soenens, 2012; Verstuyf,
Vansteenkiste, Sonenes, Boone, & Mouratidis, 2013).
The Relationship Between Changes in Psychological Need
Fulfillment and Changes in
Depressive Symptoms
Having identified significant changes in participants’
psychological need fulfillment over
the course of four assessment points, it was possible to explore
the relationship between changes
in psychological need fulfillment and changes in depressive
symptoms. Aligning with the
findings discussed above, the results indicated that increases
in autonomy, competence, and
relatedness were all significantly and negatively associated
with changes in depressive
symptoms. All three psychological needs did not significantly
differ from one another in their
associations with changes in depressive symptoms, reinforcing a
major tenet of SDT: autonomy,
competence, and relatedness are equally vital in contributing
psychological growth and health
(Deci & Ryan, 2000; Deci & Ryan, 2008; Ryan, 1995).
Moreover, this pattern of results
remained significant when accounting for the impacts of time,
neuroticism, attachment anxiety,
dysfunctional attitudes, and selective abstraction.
The present study design was not experimental, preventing any
attempt to evaluate the
causal pathways that characterize the association between
changes in psychological need
fulfillment and changes in depressive symptoms. However, the
repeated assessment of theses
-
variables of interest did allow us to explore the temporal
precedence of changes in one relative to
the other. Specifically, the present design allowed for the
construction of lagged models to
empirically determine whether shifts in one variable preceded
shifts in another. The first set of
lagged models tested whether changes in psychological need
fulfillment temporally preceded
changes in depressive symptoms. The results of this model were
significant, indicating that a
relationship was observed between participants starting to feel
more autonomous, more
competent, and more related and ensuing changes in their
depressive symptoms. The second set
of lagged models tested the reflection of this association:
whether changes in depressive
symptoms temporally preceded changes in psychological need
fulfillment. This model was also
significant; as participants experienced fewer depressive
symptoms, they subsequently began to
feel more autonomous, more competent, and more related. The
bidirectional relationship
between changes in psychological need fulfillment and depressive
symptoms is in accordance
with SDT’s concept of psychological needs as drive-inducing
resources (Deci & Ryan, 1985;
Deci & Ryan, 2000). From this perspective, psychological
need fulfillment is energizing and
mobilizes individuals to seek out additional opportunities for
psychological need fulfillment,
which further promote growth and health (Deci & Ryan,
1985).
Interestingly, the present study found that treatment group
moderated this bidirectional
temporal association. The results of the treatment group
interaction models that were constructed
suggest that the bidirectional association between changes in
psychological need fulfillment and
changes in depressive symptoms was stronger for participants who
had been randomly assigned
to cognitive therapy for treatment. Why would this temporal
relationship be stronger for
participants receiving CBT than for participants receiving ADM?
In other words, what is it about
cognitive therapy that would strengthen the effect of
psychological need fulfillment on
subsequent decreases in depressive symptoms, and vice versa,
approximately one month later?
One potential explanation is that the client-therapist
relationship may cultivate openness, self-
congruence, and self-awareness of one’s experience, defining
aspects of unified self-functioning
and the integrative process (Fournier, Di Domenico, Weststrate,
Quitasol, & Dong, 2016;
Rogers, 1963; Weinstein, Przybylski, & Ryan, 2012, 2013).
Across a myriad of traditions (see
Ryan, 1995), functioning in an integrated manner entails
enhanced access to motives, emotions,
and meanings behind one’s actions. Indeed, an integral component
of the CBT protocol used in
the present study required participants to reflect on and
evaluate their current experiences and
problems each week in the context of a thought record. In each
session, participants were
-
encouraged to reflect on their experience over the previous week
and examine thoughts,
emotions, and behaviors in a supportive therapeutic context,
with the intention of identifying
maladaptive patterns. It is possible that lagged changes in both
psychological need fulfillment
and depressive symptoms interacted with this treatment component
to catalyze participants’
tendencies toward integration, enhancing both vitality and
wellness (Ryan & Deci, 2008).
Testing the Role of Negative Cognitions in the Relationship
Between Changes in
Psychological Need Fulfillment and Depressive Symptoms
After establishing the association between changes in
psychological need fulfillment and
changes in depressive symptoms, we sought to test whether the
effect for changes in
psychological need fulfillment on changes in depressive symptoms
was mediated by changes in
negative cognitions. Contrary to our hypothesis, the indirect
effect of this mediation model
proved to be non-significant. Given that our initial power
analysis for this model had been
conducted without the complete dataset in hand, we recalculated
power for this analysis using
the updated sample size using the powerMediation package in R
(Qiu, 2015). This updated
power analysis revealed that we only had 57% power to observe
the indirect effect of our model,
ab = -2.83.
Limitations
The present study is not without limitations. First, assessment
of psychological need
fulfillment was not originally part of the larger research
project for which the participants of the
current study were recruited for. Thus, our analyses were
conducted with only nearly half of the
potential participants recruited. Moreover, participants who
completed measures of
psychological need fulfillment were significantly different from
those who did not. Although this
may be an artifact of introducing a measure much later into the
data collection process, the
participants included in the present study reported fewer
depressive symptoms, made fewer
personalization cognitive errors, and were significantly less
conscientious compared to those
who were not. Related to the limitations of the sample in the
present study, our updated power
analysis revealed that the subsample of participants who
completed indices of psychological
need fulfillment did not provide sufficient power to thoroughly
test the indirect effect of reduced
negative cognitions on the association between increases in
psychological need fulfillment and
decreases in depressive symptoms. Finally, the present study
collected data across four
-
assessment points, which varied in the length of time that
transpired between each assessment
point. Although this allows for maximal spacing for exploring
change, participants could only
end up with a maximum of four data points for most measures.
This prevented us from
constructing lagged analyses beyond a single assessment to
further explore the temporal
relationship between psychological need fulfillment and
depressive symptoms.
Future Directions
One possible future direction for research exploring the
relationship between
psychological need fulfillment and depressive symptoms concerns
exploring the various contexts
in which changes in psychological need fulfillment and changes
in depression are embedded.
Although a bidirectional relationship between changes in need
fulfillment and changes in
depressive symptoms is commensurate with SDT, it would have been
interesting to explore the
contexts in which one precedes the other. This would have been
especially interesting to explore
given the finding that that bidirectional association between
psychological need fulfillment and
depressive symptoms was stronger for participants in the
cognitive therapy condition.
Unfortunately, the present study did not employ any form of
event-contingent or diary-based
form of recording (e.g. Funder, 2016; Moskowitz, 1994) to relate
participants’ changes in
psychological need fulfillment and changes in depressive
symptoms to specific events in their
environments. Future research may benefit from exploring such
change in the context of multiple
environments. Given that increases in psychological need
fulfillment appear to have a non-
specific role in treatments for depression, it would also be
beneficial for future research to
explore how such changes in need fulfillment compare to other
non-specific factors, such as
therapeutic alliance (e.g., Gaston, 1990) and autonomous
motivation (McBride, Zuroff, Ravitz,
Koestner, Moskowitz, Quilty, & Bagby, 2010; Ryan & Deci,
2008; Zuroff, Koestner,
Moskowitz, McBride, & Bagby, 2012). Moreover, given that the
cultivation of autonomous
motivation assumes prior psychological need fulfillment, and
that autonomous motivation is an
integral part of the integrative process (Deci & Ryan, 2000;
Ryan & Deci, 2008), prospective
research should seek to confirm the causal relationship between
psychological need fulfillment
and autonomous motivation.
-
Conclusion
The present study is one of few to explore the relationship
between changes in
psychological need fulfilment and psychological symptoms,
specifically depressive symptoms.
Over the course of treatment for depression, participants
demonstrated that changes in
psychological need fulfillment have not only a reliable
trajectory within and across individuals,
but also that such changes influence one another
bidirectionally. In the present study, changes in
psychological need fulfillment were found to be associated with
changes in depressive symptoms
over an above the course of time, personality constructs (i.e.,
neuroticism and attachment
anxiety), and even negative cognitions (i.e., dysfunctional
attitudes and selective abstraction). In
this vein, the present study highlights psychological need
fulfillment as a non-specific factor
promoting healthy change across two treatment modalities. The
temporal relationship between
increases in psychological need fulfillment and decreases in
depressive symptoms was also
found to work reciprocally, reaffirming SDT’s conceptualization
of psychological needs as
drive-inducing nutrients that fuel a cycle of psychological
growth. Moreover, this reciprocal
relationship was found to be more robust in cognitive therapy
than psychopharmacology. Indeed,
when it comes to maximizing the effects of psychological need
fulfillment in the treatment for
depression, it would seem that wilted flowers flourish most when
tended with the nurturing
hands of a gardener.
-
31
References
Assor, A., & Tal, K. (2012). When parents’ affection depends
on child'