Despicable me Masselink, Maurits
DOI:
10.33612/diss.102140763
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date:
2019
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
Masselink, M. (2019). Despicable me: self-esteem and depressive symptoms among adolescents and young adults. Rijksuniversiteit Groningen. https://doi.org/10.33612/diss.102140763
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.
Download date: 27-06-2021
Self-esteem and depressive symptoms among adolescents and young adults
Maurits Willem Masselink 02-04-1989
Kampen
Cover design: Design Your Thesis | www.designyourthesis.com Layout: Design Your Thesis | www.designyourthesis.com Print: Ridderprint | www.ridderprint.nl
ISBN: 978-94-034-2144-5 ISBN (eBook): 978-94-034-2143-8
Copyright © 2019 by Maurits Willem Masselink. All rights reserved. Any unauthorized reprint or use of this material is prohibited. No part of this thesis may be reproduced, stored or transmitted in any form or by any means, without written permission of the author or, when appropriate, of the publishers of the publications.
Self-esteem and depressive symptoms among adolescents and young adults
PhD thesis
to obtain the degree of PhD at the University of Groningen
on the authority of the Rector Magnificus Prof. C. Wijmenga
and in accordance with the decision by the College of Deans.
This thesis will be defended in public on Wednesday 4 December 2019 at 12.45 hours
by
Maurits Willem Masselink
born on 2 April 1989 in Kampen
Co-supervisor Dr. E. Van Roekel
Assessment committee Prof. dr. R. Veenstra Prof. dr. J. Denissen Prof. dr. C. Finkenauer
Chapter 1. General introduction 7
Chapter 2. Domains Of Pleasure Scale (DOPS): Assessing Pleasure Across Domains
25
Chapter 3. Self-esteem in Early Adolescence as Predictor of Depressive Symptoms in Late Adolescence and Early Adulthood: The Mediating Role of Motivational and Social Factors
63
Chapter 4. The Longitudinal Association between Self-esteem and Depressive Symptoms in Adolescents: Separating Between-Person Effects from Within-Person Effects
85
Chapter 5. Dynamic Relationships Between Self-esteem, Pleasure, Sadness, and Social Experiences During Daily Life
127
Chapter 6. Reciprocal Associations Between Perceived Acceptance and Self- Esteem in Daily Life
151
Chapter 7. General discussion 167
Chapter 8. References
Nederlandse samenvatting Research Institute SHARE
Short biography and list of publications Dankwoord
195 223 231 235 239
1
CHAPTER 1
General introduction
1 1. PREFACE
Adolescence and young adulthood are developmental phases marked by major changes.
During adolescence, major psychological and physical changes occur together with major social changes (Steinberg & Morris, 2001). Adolescents change schools, have to build new social networks, work on identity formation, become involved in romantic relationships, and have to decide on study paths. During early adulthood, adults leave the parental home, have to make career choices, enter the job market or first go to college or university, and some may also start a family. All these challenges and changes can have a strong influence on how adolescents and young adults feel about themselves and their psychological well-being.
Feelings about the self and psychological well-being can also have a large impact on the choices that are made by adolescents and adults. Take the example of students deciding which study to enroll in next year. This decision will be based on their academic qualities, but is likely also based on what motivates them, and what they enjoy. Meeting up with friends, exercising, and participating in hobbies is also likely done because they give pleasure. When the motivation to do things and the experience of pleasure fall away, and sadness is experienced instead, this can severely restrict behavior and the choices that are made. Someone may stop putting effort in school work, thus affecting future career options, or someone may stop seeing friends, thereby losing a vital social network, or stop exercising, opening the risk for physical health problems.
A lack of motivation, a lack of pleasure, and feeling sad are all common features of depression (American Psychiatric Association, 2013).
Low self-esteem is another factor affecting which goals and desires are pursued. Some adolescents may want to become a pilot but incorrectly think that they are not smart enough and opt for another career choice. Others may experience pleasure in the company of others, but think that they are useless or will not be liked by others, and therefore isolate themselves instead of engaging in social interactions. Having depression and low self-esteem have been shown to have long-lasting effects on important life choices and outcomes (Kessler & Wang, 2008; Orth, Robins, & Widaman, 2012). Importantly, self-esteem and depressive symptoms have been suggested to influence each other (Sowislo & Orth, 2013).
The many challenges and choices adolescents and young adults face make them a particularly important group in the study of self-esteem and depression. In this chapter, I will further describe the concepts of depression and self-esteem and how they are proposed to be associated with each other and the social environment.
2. DEPRESSION
2.1. Depression in adolescents & young adults
Major Depressive Disorder is defined as the presence of five or more depressive symptoms for at least two weeks, of which at least one of the symptoms is either depressed mood or loss of interest or pleasure (anhedonia) (American Psychiatric Association, 2013). The other symptoms include: significant weight gain or loss, insomnia or hypersomnia, psychomotor agitation or retardation, fatigue or loss of energy, feeling worthless or excessive or inappropriate guilt, diminished ability to concentrate or indecisiveness, and recurrent thoughts of death, suicidal ideation or suicidal behavior (American Psychiatric Association, 2013).
At the beginning of adolescence, the occurrence of depression is rare (Hankin et al., 1998), with a one-year prevalence of around 1-3% at age 11. At the age of 15, the prevalence is still quite low, with a one-year prevalence of around 2.5%. However, from that age, a sharp increase in the one-year prevalence of depression occurs, and strong differences in prevalence rates between boys and girls emerge. Between the ages of 15-21 years, around 24% of girls experience symptoms qualifying depression diagnosis and 11% of boys meet the criteria for depression. The lifetime prevalence of depression for young women is around 28% at age 18 and 43% at age 21;
for young men, the lifetime prevalence is around 14% at age 18 and 21% at age 21 (Hankin et al., 1998). A more recent study shows a similar pattern, but with a sharp increase in depressive symptoms already beginning from age 11 onwards, mainly among girls (Kwong et al., 2019).
Similar findings were observed in another study which reported low lifetime prevalence of depression until about 11 years, with sharp increases in lifetime prevalence thereafter, reaching around 28% for girls and 13% for boys at age 18 (Oldehinkel & Ormel, 2015). These studies show that depression becomes highly prevalent among adolescents and young adults and that it is a high burden for those who experience depression (Gore et al., 2011).
Depression is often diagnosed using a clinical interview in which symptoms and their duration are assessed. This interview results in a binary outcome, someone is diagnosed with depression or not. However, symptoms are experienced on a continuum, both in amount of symptoms experienced and the severity of symptoms. At any given moment, the majority of the general population will not experience a full blown depression, but everybody scores somewhere on the spectrum of experiencing no to many depressive symptoms. For a more accurate understanding of depression and the factors associated with depression, it is therefore more insightful to investigate symptoms on a continuum. Moreover, experiencing depressive symptoms during adolescence has been shown to be predictive of developing full blown depression in adult life (Pine, Cohen, Cohen, & Brook, 1999; Wilcox & Anthony, 2004). Loss of pleasure (anhedonia) and feelings of worthlessness are two symptoms that have been found to be particularly predictive of depression among adolescents and adults (Gabbay et al., 2015;
Murphy et al., 2002; Pine et al., 1999; Wilcox & Anthony, 2004).
1 2.2. Assessing loss of pleasure
Loss of pleasure (anhedonia) is one of the two core symptoms of depression (the other being sad mood), and is defined as a loss of interest or pleasure during most of the day, nearly every day, in activities that used to be pleasurable (American Psychiatric Association, 2013). Anhedonia is reported by approximately 70% of people with depression (Lewinsohn, Petit, Joiner, & Seeley, 2003), and the presence of anhedonia is related to higher severity of depression and poorer treatment prognosis (Pine et al., 1999; Wilcox & Anthony, 2004). Because anhedonia is also a common symptom of schizophrenia, much of the research on pleasure experiences has been conducted on depressive or schizophrenic populations.
Pleasure experiences occur across different domains. Some of those domains relate to activities important for survival and procreation such as sensory pleasure (e.g. pleasure from food, drinks), sexual pleasure, and social pleasure (Kringelbach, 2010). Other domains of pleasure relate to, for example, monetary rewards, pleasure from pastime activities, and pleasure experiences that come from activities related to self-actualization (Kringelbach, 2010). Pleasure experiences across the domains involve shared brain circuities and neurotransmitters (Kringelbach, 2010), but the experience of pleasure and loss of pleasure can occur domain-specifically. For example, several studies have found impairments in social pleasure among schizophrenic patients, while other domains of pleasure were not or were less affected (Xie et al., 2014). There is also some evidence that depression has a stronger association with social anhedonia than with physical anhedonia (Rey, Jouvent, & Dubal, 2009).
Possibilities for easy to administer, domain-specific pleasure assessments are currently limited.
Pleasure is assessed using either laboratory tasks or questionnaires. Examples of laboratory tasks are the sweet taste tests, monetary reward tests, and affective ratings after watching video clips (Treadway, Buckholtz, Schwartzman, Lambert, & Zald, 2009; Treadway & Zald, 2011). Although laboratory tasks have the advantage of measuring pleasure or affect in the moment, they are not easy to administer and normally consist of the assessment of a single domain of pleasure.
Questionnaires are easy to administer alternatives to measure pleasure experiences, but most pleasure and anhedonia questionnaires are limited in scope as well. Some questionnaires only assess one domain (e.g. ACIPS; Gooding & Pflum, 2014) or contain items from various domains but without providing specific subscales for these domains (e.g. SHAPS; Snaith et al., 1995).
Others differentiate only between two domains (e.g. physical and social anhedonia; Chapman, Chapman, & Raulin, 1976). Moreover, several existing questionnaires contain outdated items, threatening the validity of the questionnaire, as well as the psychometric soundness. Therefore, there is a need for an easy to use questionnaire that assesses pleasure and differentiates among different domains of pleasure.
2.3. Depression and social influences
There is wide empirical support showing associations between depression, the social environment, and social functioning (Hirschfeld et al., 2000; Kupferberg, Bicks, & Hasler, 2016;
Slavich, O’Donovan, Epel, & Kemeny, 2010). Interpersonal theories of depression have highlighted the role of interpersonal difficulties in the development and endurance of depression (Coyne, 1976, 1998; Joiner, 1999). Impairments in social functioning often remain after depressive symptoms have improved (Hirschfeld et al., 2000; Kupferberg et al., 2016). Depressed people are inclined to interpret information negatively which may lead to negative social interaction outcomes (Gable & Shean, 2000). In addition, a strong focus on one’s own problems and heavy reliance on social support and reassurance from friends and relatives may negatively strain personal relationships (Evraire & Dozois, 2011; Joiner, 1999; Joiner & Metalsky, 2001). While depressed individuals may make a large appeal to their social environment, at other times they may withdraw from social interactions (Allen & Badcock, 2003). There are several reasons why social avoidance can contribute to the onset or maintenance of depression. First, social avoidance and solitude give an individual ample opportunity to ruminate about problems (Hankin, 2008), without the distraction that social interaction offers. Second, social support has been shown to be an important protective factor against developing depression (Haeffel, Voelz, & Joiner, 2007; Lee, Dickson, Conley, & Holmbeck, 2014). Avoiding social interactions leads to a decrease in opportunities to receive social support when needed. Third, social avoidance may lead to friendship dissolution, which may lead to insufficient social contact to fulfil the need to belong (Baumeister & Leary, 1995). Fourth, social interactions are vital for the development of social skills, which may be hampered when social contact is avoided. Without sufficient social skills, interactions may work out negatively (Bakker, Ormel, Lindenberg, Verhulst, & Oldehinkel, 2010), thereby reinforcing frustrations and negative feelings. Negative social interactions may ultimately lead to social rejection. Rejection by peers has been linked to adolescent psychopathology (Sentse, Lindenberg, Omvlee, Ormel, & Veenstra, 2010) and social rejection has been shown to prospectively predict depression in young adolescents (Nolan, Flynn, & Garber, 2003). Moreover, social rejection is suggested to set in motion neural and psychological processes that lead to the development of depression (Slavich et al., 2010).
Altogether, research suggests a reinforcing vicious circle between social functioning and depressive symptoms. As will be described below, self-esteem is likely involved in this vicious circle as well.
1 3. SELF-ESTEEM
3.1. Self-esteem in adolescents & young adults
Self-esteem is often defined as the global and affective evaluation of one’s own worth (Orth et al., 2012), or as the overall positive or negative affective feeling that one has about the self (Brown, Dutton, & Cook, 2001; Leary & Baumeister, 2000). This can be expressed in things like feeling good about the self, being confident, and being happy about the person one is (Bukowski & Raufelder, 2018; Nelis & Bukowski, 2019). Although self-esteem is ultimately an affective experience (Leary, 2005), it expresses itself as a combination of affective (e.g. the negative feeling after being rejected) and cognitive self-referential processes and thoughts (e.g. “I am a failure”). Self-esteem can be experienced on a trait level and a state level. Trait self-esteem relates to stable overall evaluations of the self, while state self-esteem relates to the momentary affective evaluations of the self which can fluctuate from moment to moment.
Changes in self-esteem follow developmental and social transitions. Young children tend to have overly optimistic views about the self. During middle childhood, when children enroll in primary school, they start to rely on social comparisons for judgements about the self and receive feedback about the self, leading to an “adjustment” or somewhat lower feeling of self- esteem (Harter & Whitesell, 2003). The transition into adolescence is associated with an initial sharp decrease in self-esteem, when adolescents adjust to their new social environment and physical changes, but self-esteem gradually increases during adolescence until late adulthood (Chung, Hutteman, van Aken, & Denissen, 2017; Robins, Trzesniewski, Tracy, Gosling, & Potter, 2002; von Soest et al., 2016). Although individuals’ personal self-esteem levels change over time, differences in self-esteem between individuals remain relatively stable, although less so during adolescence (Chung et al., 2017; Donnellan, Kenny, Trzesniewski, Lucas, & Conger, 2012; Robins et al., 2002). This stability in self-esteem differences between individuals means that someone with a higher-than-average self-esteem at age 10 is likely to have higher than average self- esteem at age 20, despite changes in absolute self-esteem level.
Self-esteem has received much empirical attention because it is seen as an important motivating factor (Deci & Ryan, 1991, 1995; Hogg, Abrams, Otten, & Hinkle, 2004; Ryan & Deci, 2000) and it is associated with many life outcomes. Humans are assumed to have a motivation to have and maintain high self-esteem and to feel good about the self (Baumeister, Tice, & Hutton, 1989; Harter, 1993; Heimpel, Elliot, & Wood, 2006). Positive views about the self have been suggested to be so important that humans are willing to distort reality and adopt overly positive views about the self (Taylor & Brown, 1988), and interpret information in a self-serving manner (Campbell & Sedikides, 1999). High self-esteem is further associated with reward approach behavior (Erdle & Rushton, 2010; Kuppens & Van Mechelen, 2007; Park, 2010; Tice & Masicampo, 2015). Individuals with high self-esteem feel confident to take up tasks, explore new things, socialize with others, and look for opportunities to further increase their self-esteem. Moreover,
when faced with failure and negative mood, individuals with high self-esteem are motivated to restore their mood, while individuals with low self-esteem are less motivated to do so (Heimpel, Wood, Marshall, & Brown, 2002). Low self-esteem is associated with a focus on avoiding failure and rejection, which expresses itself in fewer social interactions, giving up on goals earlier, and displaying more avoidance behavior compared to those with higher self-esteem (Erdle & Rushton, 2010; Kuppens & Van Mechelen, 2007; Park, 2010). Having high self-esteem has been associated with more positive life outcomes compared to having low self-esteem. Outcomes include, for example, higher rates of college graduation, higher job satisfaction and security, higher income, less criminal behavior, and better physical and psychological health (Trzesniewski et al., 2006).
3.2. Self-esteem & social influences
Several factors are likely to influence self-esteem. According to William James, self-esteem is determined by the ratio of one’s aspirations in domains that one considers important, and the success one has in those domains (James, 1890). Thus, failing in sport will only negatively affect self-esteem in those who think it is important to be a good athlete. However, failing or succeeding only has meaning when the outcome is compared to some social norm or the performance of other people. That is why most scholars have described self-esteem formation from a social framework. Central to many of the views on self-esteem, is that the social environment influences how people evaluate themselves, that people are generally motivated to keep that evaluation as favorable as possible, and that the outcome of the evaluation is what represents self-esteem. According to the looking-glass-self perspective, people actively try to infer how they are perceived by significant others (Cooley, 1902) such as family members, friends, classmates, and teachers. Significant others serve as a social mirror, in which we look to get information about ourselves. The perceived opinions of multiple significant others may be averaged into an overall generalized other (Mead, 1925, 1934). If one infers that others think positively about oneself, self-esteem will be high, whereas if one infers that others think negatively about oneself, self-esteem will be low. The role of the social environment in the formation of self-concepts and self-esteem is also a central point in the attachment literature. Children of parents who are available and responsive to the child’s needs form rudimentary images of the self as positive and deserving of love, while children of unresponsive parents form negative images about the self (Laible, Carlo, & Roesch, 2004). Children and adolescents with secure attachments to parents and peers have been shown to have higher levels of self-esteem than children with insecure attachments (Arbona & Power, 2003; Laible et al., 2004). Another indication of the influence of the social environment on self-esteem is the temporal drop in self-esteem during transitions to new social environments, such as when changing schools (Harter, 1993; Harter & Whitesell, 2003).
1 An influential theory of self-esteem within the social framework, sociometer theory, has a
radically different conceptualization of self-esteem (Leary, 2005; Leary, Terdal, Tambor, & Downs, 1995). Coming from an evolutionary approach, the starting point of sociometer theory is that our ancestors had higher chances of survival when in groups than when alone, eventually leading to the development of a need to belong (Baumeister & Leary, 1995). Sociometer theory conceptualizes self-esteem as a warning system that monitors the risk of social rejection.
Self-esteem reflects the perceived relational value of an individual. Relational value indicates the degree someone perceives his or her relationship with other individuals as valuable and important (Leary, 2005). The lower the perceived relational value, the more likely someone is to be rejected by others, whereas high relational value is an indication that one will be or is included by others. Self-esteem is thus quite literally seen as a sociometer. According to this theory, rather than having a motivation to have high self-esteem, people have a motivation to have high relational value. Low self-esteem is an alarm signal that motivates behavior to increase relational value. The authors of sociometer theory compare the sociometer to the fuel meter of a car, and the feeling of self-esteem as the dial pointing to the fuel level. Just like the driver of a car is not motivated to have the fuel dial pointing to “Full”, but rather to avoid a lack of fuel, humans are not motivated to keep self-esteem high, but rather to keep relational value high.
Despite differences in theoretical models and conceptualizations of self-esteem, there are similarities in how the social environment is expected to influence self-esteem. All models predict that failure, negative social evaluations, and social rejection negatively influence self- esteem. Similarly, success, positive feedback, and social acceptance are theorized to predict increases in self-esteem. For other aspects of the associations between self-esteem and the social environment, it is less straightforward what to expect, for example with regard to interaction between state and trait self-esteem and social experiences. Effects of social acceptance and social rejection on self-esteem may be moderated by trait self-esteem. Those with high trait self-esteem may not respond as strongly to acceptance and rejection as those with lower trait self-esteem. Another open question remains whether social acceptance and social rejection exert the same influence on self-esteem. From a sociometer perspective, the warning system (i.e., self-esteem) should be especially attuned to signals of danger. However, a meta-analysis on the effects of social acceptance and rejection on self-esteem suggests exactly the opposite, namely that acceptance was associated with an increase in self-esteem, while rejection was not significantly associated with a decrease in self-esteem (Blackhart, Nelson, Knowles, & Baumeister, 2009). This meta-analysis was based on laboratory studies, so it is unknown if and how the results generalize to settings outside of the laboratory.
The different conceptualizations of self-esteem have different predictions on how self-esteem affects future social behavior and preferences. On the one hand, low self-esteem is associated with avoidance motivation and behavior, so one would predict that individuals with low self- esteem have a lower desire for social contact and less social contact than those with high self-
esteem. On the other hand, sociometer theory suggests that after experiencing low self-esteem, individuals should be motivated to restore self-esteem, and thus actively seek out opportunities to do so. Although there are exceptions (e.g., Denissen, Geenen, van Aken, Gosling, & Potter, 2008; Murray, Griffin, Rose, & Bellavia, 2003), most studies investigating associations between social factors and self-esteem are correlational or laboratory studies. Much remains unknown about how the social environment and self-esteem reciprocally influence each other.
4. DEPRESSION & SELF-ESTEEM
4.1. The association between depressive symptoms and self-esteem
An abundant amount of research has shown that self-esteem and depressive symptoms co- occur among adolescents (see Sowislo & Orth, 2013 for a review). Before discussing theoretical models describing this association, it is important to address the issue of conceptual overlap between self-esteem and depression. Worthlessness is one of the symptoms of depression and an indication of low self-esteem. Due to this overlap and the strong positive correlations often found between the constructs, it has been argued that self-esteem and depression are essentially the same construct (Watson, Suls, & Haig, 2002). Several things speak against this conclusion. The meta-analyses of Sowislo and Orth (2013) found an average correlation of .57 between self-esteem and depressive symptoms. Although this indicates a strong correlation, it is not strong enough to suggest equivalence. Other evidence that speaks against equivalence is the higher stability of self-esteem than depressive symptoms (Orth, Robins, & Roberts, 2008); the finding that associations between self-esteem and depressive symptoms are still consistently found when construct overlap is taken into account using longitudinal studies (Sowislo & Orth, 2013); and that common factor models tend to have poor model fit (Orth et al., 2008; Rieger, Göllner, Trautwein, & Roberts, 2016). This indicates that self-esteem and depressive symptoms are related but largely independent constructs.
The two most extensively investigated models describing the association between self- esteem and depressive symptoms both assume at least partial independence between self- esteem and depressions. In line with several theories of depression (Abramson & Metalsky, 1989; Beck, 1967, 2008), the vulnerability model states that self-esteem is not only a symptom of depression, but it precedes the development of other depressive symptoms. Having low self- esteem is seen as a vulnerability factor for developing depressive symptoms, more so than the other symptoms of depression. The scar model states that going through a depressive episode is detrimental for self-esteem, leaving a psychological scar on self-esteem that remains even after remission of the depressive episode. The empirical evidence provides evidence for both the vulnerability and the scar model (Sowislo & Orth, 2013; Steiger, Fend, & Allemand, 2015).
Most support has been found for the vulnerability model, both with regard to presence of
1 associations and predominance of effects (i.e. stronger effects from self-esteem to depressive
symptoms than the other way around). Vulnerability effects are, on average, twice as large as scar effects (Sowislo & Orth, 2013). An interesting question is whether having low self-esteem during early adolescence is an enduring vulnerability for developing depression later in life or if this vulnerability disappears over time. Currently, only a couple of studies have looked into this question. Two studies suggest that self-esteem during adolescence remains a vulnerability factor for developing depression up to 23 years later (Steiger, Allemand, Robins, & Fend, 2014;
Trzesniewski et al., 2006). Another study found no significant association between self-esteem measured at age 15 and depressive symptoms at age 25 (Boden, Fergusson, & Horwood, 2008).
More research on this topic is needed and would benefit from investigating the mediators that underlie the longitudinal association between self-esteem and depressive symptoms.
The vulnerability and scar models provide rudimentary models for a likely more complex underlying process. Although there may be direct effects between self-esteem and depressive symptoms, the associations are likely mediated by other variables (Kuster, Orth, & Meier, 2012; Orth, Robins, Meier, & Conger, 2016). Considering their central role in both self-esteem and depression, motivational patterns and the social environment are important factors to integrate in models describing the associations between self-esteem and depressive symptoms.
Importantly, approach and avoidance motivation and social factors may not only serve as mediators between self-esteem and depressive symptoms, but these variables may possibly be part of a complex system with reciprocal associations between all variables. If we want to gain insight into this system, it will be necessary to investigate them together in the same model.
Although several parts of the associations between self-esteem, depressive symptoms, and social factors have been investigated, they have not been investigated simultaneously.
4.2. Associations across different time frames
Repeated measures designs can provide insight into the possible mechanisms underlying the associations under investigation. Because of the temporal spacing between measures, it is possible to investigate the direction of effects (e.g. does self-esteem precede depression or the other way around), as well as directional dominance (e.g. is the effect of self-esteem on depression stronger than the reversed association).
When investigating temporal associations, it is important to consider the time frame in which associations are expected to occur. Inferences about processes measured on a short time frame may not apply to processes at larger time frames, because the mechanism may differ across time frames. Keijsers and van Roekel (2018) give the example of investigating the walk and gallop of a horse. The mechanism a horse uses to walk (e.g. the order of leg movement, the way legs are extended and retracted, etc.) differs substantially from the mechanism used when in gallop, it is not just a matter of speeding up. Investigating the mechanism of the walk won’t tell you anything about the mechanism of the gallop and vice versa because they are
qualitatively different processes. Keijsers and van Roekel coined the term “galloping-horse- fallacy” to describe the situation where researchers make inferences about processes on one time frame based on processes measured at a different time frame. The example of the walk and gallop of a horse also shows that if we want to understand the movement of a horse, we will have to investigate it across the multiple relevant time frames. Psychological processes are part of a complex system simultaneously operating at different time frames. Thus, if we want to understand the mechanisms underlying the associations between self-esteem and depressive symptoms, these associations have to be investigated on multiple time frames. Equally important is the use of research designs and statistical techniques that are appropriate for the time frame under consideration.
Many studies have investigated the association between self-esteem and depressive symptoms over time. Most of those studies employed designs with two to four assessments per individual, with regular time intervals between assessments ranging from weeks to a couple of years, with sometimes a follow-up assessment up to 23 years later (Sowislo & Orth, 2013;
Steiger et al., 2014; Trzesniewski et al., 2006). These kind of designs are useful to investigate trait like associations between self-esteem and depressive symptoms. They can also be insightful to investigate whether there are long-lasting effects, or to investigate whether associations remain stable across developmental periods. However, longitudinal designs with long time intervals between assessments may not be able to detect some of the effects because these effects have faded away long before the constructs are assessed again.
Many psychological processes are likely to occur and influence each other on a small time scale, from day to day, or from moment to moment during the day. Experience Sampling Methodology (ESM; Larson & Csikszentmihalyi, 1983) or Ecological Momentary Assessment (EMA; Shiffman, Stone, & Hufford, 2008) are methods to investigate associations on the small time scales that many psychological processes are proposed to occur. ESM is an intensive data collection design in which participants complete assessments daily or multiple times per day for several days or weeks, nowadays mostly using smartphones to register their responses. ESM designs have the advantage of large ecological validity because they provide the opportunity to measure constructs in the moment, therefore reducing recall bias. A small number of ESM studies have included either self-esteem, depressive symptoms, or social variables, or a combination of two of these factors (e.g., Brown, Strauman, Barrantes-Vidal, Silvia, & Kwapil, 2011; Clasen, Fisher,
& Beevers, 2015; Denissen, Penke, Schmitt, & van Aken, 2008; Nezlek & Plesko, 2003), but to the best of my knowledge these three factors have never been studied together. It therefore remains unclear how self-esteem, depressive symptoms, and social factors (i.e., acceptance, rejection, social contact, social motivation) affect each other during daily life.
1 4.3. Between-person and within-person effects
When investigating psychological temporal associations and mechanisms, we are usually interested in within-person associations, and sometimes whether those associations are moderated by between-person moderators. For example, do changes in state self-esteem lead to changes in state depressive symptoms and are those associations moderated by trait self-esteem? To answer these kinds of research questions, statistical analyses that are able to separate within-person and between-person effects are required.
Failure to use the appropriate techniques can lead to uninterpretable results and conclusions that may even be in the opposite direction of the true effect (Fisher, Medaglia, & Jeronimus, 2018; Hamaker, Kuiper, & Grasman, 2015; Kievit, Frankenhuis, Waldorp, & Borsboom, 2013).
Statistical techniques that do not separate within-person and between-person effects only provide accurate results about within-person processes under very strict conditions, known as ergodicity. Ergodicity is only present when cross-sectional population level parameters, which include means, variances, and covariances, are identical to the same within-person parameters over time (Hamaker, 2012; Molenaar, 2004; Molenaar & Campbell, 2009). This would imply, for example, that there is no developmental process in self-esteem, that all individuals have the same average self-esteem level and the same variance in self-esteem over time, and that the associations between self-esteem and depressive symptoms would be the same for all individuals (Fisher et al., 2018; Hamaker, 2012). It is clear that the conditions of ergodicity are unlikely to hold in psychological research and, therefore, that statistical techniques capable of separating within-person effects from between-person effects are needed. Unfortunately, a substantial body of research investigating the association between self-esteem and depressive symptoms is based on a method of analysis that does not separate within-person effects from between-person effects (Hamaker et al., 2015; Masselink, van Roekel, et al., 2018). This questions the validity of a substantial part of existing literature that has examined the association between self-esteem and depressive symptoms. Therefore, studies using methodologies capable of separating within-person effects from between-person effects are much needed in order to get a better understanding of the association between self-esteem and depressive symptoms.
5. THIS THESIS
5.1. Outline of the remainder of this thesis
Chapter 2 addresses the limitations of existing pleasure and anhedonia measures with the development of the Domains Of Pleasure Scale (DOPS). The DOPS is a pleasure questionnaire designed to assess and distinguish between pleasure across the domains of social pleasure, sexual pleasure, perceptual pleasure, and pleasure derived from personal achievements. We extensively tested the validity and psychometric properties of the DOPS.
Considering the limited and mixed evidence for self-esteem as an enduring vulnerability factor for developing depressive symptoms over many years, Chapter 3 describes whether self- esteem at early adolescence is an enduring vulnerability factor for developing depression in late adolescence and early adulthood. In addition, we investigated whether longitudinal associations between self-esteem and depressive symptoms were mediated by approach and avoidance motivation and social factors (i.e. social problems, social support and social contact).
Chapter 4 addresses the issue that previous studies investigating the association between self-esteem and depressive symptoms used a method of analysis that does not separate within- person effects from between-person effects. Using a method that does separate within-person effects from between-person effects, we investigated the longitudinal associations between self-esteem and depressive symptoms in three datasets, covering early adolescence to early adulthood.
As described in previous sections, self-esteem, depressive symptoms, and social factors are likely part of a complex process influencing each other during the day. In the study described in Chapter 5, we aimed to elucidate part of this process by examining reciprocal associations between self-esteem, depressed mood, and social factors using two ESM studies. It was investigated whether self-esteem was associated with sadness and pleasure, and whether self- esteem, sadness, and pleasure were associated with the amount of social contact and social motivation.
In Chapter 6, an ESM design was used to investigate hypotheses that follow from sociometer theory, namely that social acceptance and social rejection affect self-esteem. We further explored whether low self-esteem predicted social acceptance and social rejection. In addition, we investigated whether the sociometer of individuals with low trait self-esteem was more sensitive than those with high trait self-esteem.
Chapter 7 provides the general discussion in which the insights across these studies are synthesized, and the implications of findings, together with challenges and improvements of science, are discussed.
1 5.2. Datasets used
In this dissertation I used several different datasets covering early adolescence to adulthood.
Cross-sectional datasets. In Chapter 2 we used three cross-sectional datasets to develop and test the DOPS. The No Fun No Glory study was set up to investigate anhedonia among young adults. As part of this study, 2,937 adults ages 18-24 years filled in the DOPS between February and April 2015. Between March and April 2017, a subset of 962 individuals participated in a follow-up study to further validate the DOPS. In this follow-up study we also included a sample of 225 first year psychology students from Tilburg University.
Longitudinal datasets. In Chapter 3, we used data from four measurement waves (T1- T3, T5) of the large, prospective cohort study Tracking Adolescents’ Individual Lives Survey (TRAILS). Data were collected between 2001 and 2013, with 2-3 year time intervals between measurement waves. Data collecting started during early adolescence (T1 N = 2,128, ages 10-12 years) and ended in early adulthood (T5, ages 21-24 years).
In Chapter 4, we used three different longitudinal datasets, together spanning a developmental period from early adolescence to early adulthood. The data of Study 1 were collected between 2013 and 2015, with three measurement waves spaced one year apart. T1 data were collected among 1,223 Dutch first grade secondary school students (average age 12.8 years). Due to a drop-in design, a total of 1,948 adolescents participated at T3 (average age 15.4 years). In Study 2, we used data from the first three waves of a Belgium study with one year time intervals between measurements. Data collection took part in schools between 2009 and 2011. The average age at T1 was 15.8 years and at T3 was 17.4 years. A total of 1,455 participants took part in the study which had a drop-in design like Study 1. Study 3 consisted of an American sample (N = 316) collected at Rutgers University with 1.5 year time intervals between measurements. Data were collected between 2008 and 2014. The average age was 11.5 years at T1 and 15 years at T3.
ESM datasets. In Chapter 5, we used two ESM datasets. In Study 1, data came from the ESM part of the No Fun No Glory data. Out of the 2,937 young adults screened for anhedonia, 69 anhedonic participants and 69 matched control participants were enrolled in an ESM study.
We only used the data of the control group, which consisted of young adults who scored higher than average on a pleasure measure. Data from the first 30 days of ESM data collection were used, with 3 assessments per day on fixed six-hour time intervals. In Study 2, ESM data came from the HowNutsAreTheDutch (HND) study. The HND sample was recruited from the general population of the Netherlands by a crowdsourcing procedure. The sample consisted of 938 individuals (average age 38.8 years) who provided data for 30 days, three times per day, on fixed six-hour time intervals. In Chapter 6, we used an ESM dataset collected among 228 college students of Tilburg University. The average age of the participants was 19.4 years. Participants completed five questionnaires per day at semi-random time intervals for 11 consecutive days.
Overview of datasets used DatasetDesignData collectionSampleChapter No Fun No Glory - ScreeningCross-sectionalParticipants filled in a battery of questionnaires including the DOPSN = 2,937 age 18-242 No Fun No Glory - DOPS validationtest - retest designBaseline: DOPS + validation questionnaires Retest: 14 days laterN = 962 (149 retest) age 19-282 Tilburg University - DOPS validationtest - retest designBaseline: DOPS + validation questionnaires Retest: 14 days laterN = 225 (194 retest) age 17-212 No Fun No Glory - ESMExperience sampling30 days 3 measures per day fixed 6-hour time intervals
N = 69 age 18-245 HowNutsAreTheDutchExperience sampling30 days 3 measures per day fixed 6-hour time intervals
N = 938 mean age 38.85 ESM Tilburg UniversityExperience sampling11 days 5 measures a day semi-random intervals
N = 228 mean age 19.46 US datasetLongitudinalThree 1.5 year interval wavesN = 316 mean age at T1 11.54 Dutch dataset Longitudinalthree annual wavesN = 1,948 mean age at T1 12.84 Belgium dataset Longitudinalthree annual wavesN = 1,455 mean age at T1 15.84 TRAILSLongitudinalT1-T3, T5N = 2,128 age 10-12 at T13
2
CHAPTER 2
Domains Of Pleasure Scale (DOPS):
Assessing Pleasure Across Domains
Masselink, M.
van Roekel, E.
Heininga, V. E.
Vrijen, C.
Oldehinkel, A.
Under review
ABSTRACT
The Domains Of Pleasure Scale (DOPS) is a newly developed questionnaire designed to measure the multifaceted aspects of pleasure. It assesses levels of pleasure across different domains (e.g., social, physical). The psychometric properties of the DOPS were tested in two studies (Study 1: N = 2937, Mage = 21.4 years, SD = 1.9; Study 2: N = 1187, Mage = 22.84, SD = 2.23). In line with the multifaceted view of pleasure experiences, comparisons with a one factor solution showed that pleasure may be best investigated domain-specific rather than aggregated over all items.
Across studies, exploratory and confirmatory factor analysis showed that pleasure was reliably measured in the domains of social pleasure, sexual pleasure, perceptual pleasure, and pleasure in personal achievements. Measurement invariance was established across sex and educational level in Study 1, but not in Study 2. The DOPS is a promising instrument to measure pleasure across different domains and may help researchers and clinicians to pinpoint interventions.
Keywords: Domains Of Pleasure Scale (DOPS), consummatory pleasure, loss of pleasure, anhedonia
2 The ability to experience pleasure is important for human functioning and well-being. The
decrease or inability to experience pleasure, also called anhedonia, is common in different sorts of psychopathology, for example schizophrenia and substance use, but most prominently in depression, where it is one of the two core symptoms (American Psychiatric Association, 2013;
Blanchard & Cohen, 2006; Franken, Rassin, & Muris, 2007). Sixty-five to 77% of people with a Major Depressive Disorder report anhedonia (65-77%; Buchwald & Rudick-Davis, 1993; Lewinsohn, Petit, Joiner, & Seeley, 2003; Roberts, Lewinsohn, & Seeley, 1995).
Pleasure can be experienced in different domains. A common distinction is between pleasure derived from physical experiences (e.g. touch, smell, hearing) and pleasure derived from social experiences (Kringelbach, 2010). Other, partly overlapping, domains are sexual pleasure, appetitive pleasure, pleasure from pastime activities and pleasure experiences that come from activities related to self-actualization (Berridge & Robinson, 1998; Kennedy, Dickens, Eisfeld, &
Bagby, 1999; Rizvi et al., 2015; Snaith et al., 1995). Although all hedonic pleasure experiences involve shared brain circuities and neurotransmitters (Kringelbach, 2010), research on loss of pleasure has shown that pleasure and loss of pleasure may occur domain-specific. For example, schizophrenic patients with social anhedonia have decreased pleasure responses towards social rewards compared to controls, but do not differ in their response towards monetary rewards (Xie et al., 2014). Furthermore, physical anhedonia has been related to attenuated responses to pleasure-inducing stimuli (Blanchard, Bellack, & Mueser, 1994; Fitzgibbons & Simonst, 1992;
Kaviani et al., 2004; but see e.g. Berenbaum & Oltmanns, 1992; Berenbaum, Snowhite, & Oltmanns, 1987), while social anhedonia may not affect pleasure responses to affective stimuli (Blanchard et al., 1994; but see e.g. Dowd & Barch, 2010).Together these findings suggest that physical and social anhedonia constitute partly different kinds of impairments. In sum, evidence is supportive of the notion that pleasure experiences may occur domain-specific.
If pleasure experiences can occur domain-specific, it would benefit researchers and clinicians to be able to measure pleasure domain-specific. We present the Domains Of Pleasure Scale (DOPS) which was developed to measure domain-specific pleasure experiences as well as general pleasure experiences. The DOPS is designed to measure pleasure in the most researched area of pleasure, consummatory pleasure, which refers to pleasure experienced during the activity itself (as opposed to motivational and anticipatory pleasure). In the remainder of this introduction, we briefly introduce the currently existing pleasure or loss of pleasure questionnaires. Next, we elaborate on what the DOPS adds to existing questionnaires. We focus on self-report questionnaires because of their ease of administering, and will leave experimental or interview based assessment methods aside.
Existing Pleasure and Anhedonia Questionnaires
To date, the most frequently cited anhedonia and pleasure questionnaires are the 61-item Revised Physical Anhedonia Scale (RPAS; Chapman & Chapman, 1978) and the 40-item Revised Social Anhedonia Scale (RSAS; Eckblad, Chapman, Chapman, & Mishlove, 1982). The two Chapman scales were originally developed to measure social and physical anhedonia in schizophrenia patients by posing descriptive items, presented in a true/false response format. Other often-cited scales include the Fawcett-Clark Pleasure Scale (FCPS; Fawcett, Clark, Scheftner, & Gibbons, 1983), the Snaith-Hamilton Pleasure Scale (SHAPS; Snaith et al., 1995), and The Temporal Experiences of Pleasure Scale (TEPS, Gard, Gard, Kring, & John, 2006). The FCPS is a 36-item pleasure scale that asks participants to imagine how much pleasure they could experience in a variety of situations regardless of the real life applicability of the situation. The SHAPS is a 14-item questionnaire developed as an easy and quick instrument to measure anhedonia. The TEPS consists of a 10- item anticipatory pleasure and 8-item consummatory pleasure scale. More recently developed questionnaires are the Anticipatory and Consummatory Interpersonal Pleasure Scale (ACIPS;
Gooding & Pflum, 2014), the Dimensional Anhedonia Rating Scale (DARS; Rizvi et al., 2015), the Motivation and Pleasure Scale – Self-Report (MAP-SR; Llerena et al., 2013), the Specific Loss of Interest and Pleasure Scale (SLIPS; Winer, Veilleux, & Ginger, 2014), and the Leuven Affect and Pleasure Scale (LAPS; Demyttenaere, Mortier, Kiekens, & Bruffaerts, 2017). Table 1 depicts the domains of pleasure that are covered by the above mentioned questionnaires. Due to their scope and content, these questionnaires share one or more of the below-described limitations.
The main issue with existing questionnaires is that, with the exception of the DARS and Chapman scales, they do not differentiate between domains of pleasure. The Chapman scales only cover physical and social pleasure and consist of a large number of items, making them potentially burdensome to fill in. Other questionnaires contain items mostly aimed at a single domain, like perceptual pleasure (TEPS) or social pleasure (ACIPS). The SHAPS and the SLIPS do cover multiple domains, but do not differentiate between them by means of subscales. The DARS is the exception with subscales covering multiple domains. A disadvantage of the DARS is that its domain scores represent a mix of consummatory, motivational and anticipatory pleasure, making it unclear whether scores around the middle of the scale represent overall moderate pleasure across consummatory, motivational and anticipatory pleasure, or high scores on one dimension and low scores on another. Additional issues concern the item quality and appropriateness. The FCPS, RPAS, RSAS and the TEPS contain several items that may be considered outdated (e.g.
items about organ music; RPAS), are possibly not pleasurable to many people (e.g. “I love it when people play with my hair”; TEPS), or are only relevant for specific cultural or age groups (e.g. items about snow, TEPS, RPAS and FCPS; items about sex, RPAS, FCPS). To assure wide validity, pleasure measures should contain items that are appropriate for all individuals and conditions and should contain items which are relevant to the majority of people.
2
TABLE 1. The Domains, Level and Change Of Pleasure or Anhedonia Measured by Existing Questionnaires
RPAS RSAS FCPS SHAPS TEPS ACIPS DARS MAP-SR SLIPS LAPS
Domains
Social ü ü ü ü ü ü ü
Physical / Perceptual ü ü ü ü ü ü ü ü
Other ü ü ü ü ü
Differentiates domains ü
Level of pleasure/anhedonia measured ü ü ü ü ü ü ü ü ü
Change in pleasure measured ü
Note. RPAS = Revised Physical Anhedonia Scale; RSAS = Revised Social Anhedonia Scale; FCPS = Fawcett-Clark Pleasure Scale;
TEPS = Temporal Experiences of Pleasure Scale; ACIPS = Anticipatory and Consummatory Interpersonal Pleasure Scale; DARS = Dimensional Anhedonia Rating Scale; MAP-SR = Motivation And Pleasure Scale-Self-Report; SLIPS = Specific Loss of Interest and Pleasure Scale; LAPS = Leuven Affect and Pleasure Scale
The DOPS
We have highlighted the need to conceptualize pleasure as a multifaceted construct, and argued that there are currently no pleasure questionnaires available that sufficiently differentiate between the various domains of consummatory pleasure. The Domains Of Pleasure Scale (DOPS) was developed to differentiate between physical pleasure, social pleasure, sexual pleasure, intellectual activities or esthetic activities, and pleasure derived from hobbies. Additional features of the DOPS reported in the Supplemental materials are the assessment of change in pleasure, which can be used to determine loss of pleasure, and the duration and possible cause of loss of pleasure, if any.
The DOPS was initially assessed in a small sample, and, after improvements, assessed in two larger community samples of young adults. Adolescents and young adults are a highly relevant population in this respect, as research suggests that around 20% of adolescents and young adults experience periods of anhedonia (Bennik, Nederhof, Ormel, & Oldehinkel, 2014).
We examined (a) the factor structure and reliability of the DOPS, (b) measurement invariance, (c) the interrelations of the DOPS measures, (d) possible sex and educational differences, (e) test- retest reliability, and (f) associations of the DOPS with other pleasure, anhedonia and depression measures. All data, syntaxes and output of the reported studies are made available via the Open Science Framework (https://osf.io/gcn3d/), as is the DOPS itself.
STUDY 1 METHODS
Scale Development
Originally the DOPS was designed to measure not only consummatory pleasure, but motivational and anticipatory pleasure as well. Since the motivational and anticipatory measures appeared to require further development, we chose to exclude those items from the DOPS and to omit their description here. We shortly elaborate on this decision further in the discussion and provide information about the motivational and anticipatory items via the OSF page (https://osf.io/
gcn3d/).
Item selection was an iterative process in which we first identified domains of pleasure that were explicitly or implicitly mentioned in the relevant literature, that is, physical, social, sexual, achievement and self-actualization, and esthetic pleasure. We inspected existing (at that time) pleasure and anhedonia scales (RPAS, RSAS, TEPS, SHAPS and SLIPS) for suitable items, and adjusted and created new items to cover all domains. This resulted in an item pool of 180 items, which was evaluated on suitability by the authors of the present article. In order to be judged as suitable, consensus was needed among the authors that an item was contemporary and suitable to all ages and conditions, and that it was likely to reflect experiences that are considered pleasurable by the majority of people.
This screening of possible suitable items resulted in a preliminary selection of 26 pleasure experiences, covering the five before-mentioned domains. Responses concerned pleasure experiences during the past two weeks, and had to be indicated on a Visual Analogue Scale (VAS) with “Not at all” and “Very much” as left and right anchors. The anchors corresponded with the values 0 and 100 respectively. Due to its wide range, a VAS has the advantage of allowing to detect subtle individual differences and intra-individual changes after repeated measures. In addition, we asked participants to rate their overall, domain-overarching pleasure. For this item, the anchors on the VAS scale were “I hardly ever enjoy myself” and “I almost always enjoy myself”.
This yielded a total of 27 pleasure items.
This initial version of the DOPS was piloted in a sample of 273 intermediate vocational educational and university students (mean age = 19.6 years, SD = 3.2, 68% female). Twenty-one specific pleasure items were retained after discarding items concerning situations that were not generally considered as pleasurable (i.e., mean pleasure score below 60; five items), combining two overlapping items, and adding one new social item (I enjoy being with family or good friends). The discarded items were about (1) feeling a soft carpet, beach or grass on bare feet, (2) hearing certain sounds (e.g., singing birds, leaves rustling), (3) activities requiring light exercise (e.g. going for a walk, yoga), (4) reading a book or magazine, and (5) activities that require hard thinking (e.g. chess, studying, puzzles, and challenging computer games). The combined item
