Feeling Fat and Depressed?

Self‐concept (SC) is an individual's self‐perception of his or her attributes, deficiencies, relationships, and values (Sánchez & Roda, 2007). It is a multidimensional construct related as expected to both positive phenomena (e.g., well‐being, positive coping with adversity; Marsh, 2007) and negative phenomena (e.g., greater anxiety and depression, lower overall health). Factor analyses of various SC measures indicate that SC has first‐order factors representing specific aspects of SC (e.g., academic ability, physical appearance) and a second‐order “global” SC factor that approximates self‐esteem (e.g., Dishman et al., 2006; Fitts & Warren, 1996; Kaminski, Schafer, Neumann, & Ramos, 2005). Physical SC (PSC) represents the level of contentment one has regarding one's appearance, body shape, and weight. Decades of research with women and girls indicate that low PSC is consistently associated with constructs such as greater weight concern, lower self‐esteem, disordered eating, and depression (e.g., Friedman et al., 2005; Ting, Huang, Tu, & Chien, 2012).

When researchers began to study the correlates and consequences of low PSC in men (e.g., Brand, Rothblum, & Solomon, 1992; Herzog, Newman, & Warshaw, 1991; Siever, 1994), they discovered that the variety of ways men can be dissatisfied with their bodies made them more challenging to study than women. For example, men report dissatisfaction by believing they are too thin (de Wit, van Straten, van Herten, Penninx, & Cuijpers, 2010), too fat (Grogan & Richards, 2002; Tylka, 2011), too weak (Kaminski & Hayslip, 2006), or not muscular enough (i.e., muscle dysmorphic disorder [MDD; Pope, Katz, & Hudson, 1993]; McFarland & Kaminski, 2009; Tylka, 2011). Researchers need to attend to this heterogeneity of body dissatisfaction when investigating low PSC among men, because there is no evidence that the risk factors, correlates, and outcomes of low PSC due to dissatisfaction with thinness versus fatness versus lack of muscularity are the same (e.g., Grossbard, Neighbors, & Larimer, 2011; Kaminski & McFarland, 2006; Tylka, 2011). Thus, we created a relatively homogeneous sample of men in that all dissatisfaction with PSC was due to real or perceived adiposity.

The vast majority of the extant literature on PSC and mental health outcomes in college men focuses on those who have MDD or bulimia (e.g., Grieve, 2007; Olivardia, Pope, Borowiecki, & Cohane, 2004). These conditions have low prevalence among undergraduates (Eisenberg, Nicklett, Roeder, & Kirz, 2011; Kaminski & McFarland, 2006). Overweight and obesity, however, are becoming increasingly prevalent among college men (Pedrelli, Nyer, Yeung, Zulauf, & Wilens, 2015), with base rates of one third and higher (Nagata, Garber, Tabler, Murrary, & Bibbins‐Domingo, 2018; Odlaug et al., 2015). Being overweight or obese is associated with body dissatisfaction, low PSC, and depression in young men (e.g., Calzo et al., 2012; Odlaug et al., 2015; Suchert, Hanewinkel, & Isensee, 2016), and poor PSC independently predicts higher depressive symptoms (Richard, Rohrmann, Lohse, & Eichholzer, 2016).

Although meta‐analyses of longitudinal research studies suggest the relationship between poor PSC and depression is bidirectional (e.g., Luppino et al., 2010; Pereira‐Miranda, Costa, Queiroz, Pereira-Santos, & Santana, 2017; Rooke & Thorsteinsson, 2008), the extant literature examining the mechanisms by which having low PSC leads to depression has typically focused on groups of girls and women who are at a life stage when body changes are normative—for example, pubescent, pregnant, menopausal (e.g., Kim & Moon, 2006; Riquin et al., 2019; Siegel, 2002). Men with low PSC attributed to perceived or real overweight (due to adiposity), however, are also at increased risk for depression (e.g., Blashill, 2010; Calzo et al., 2012; Luppino et al., 2010), and understanding factors or mechanisms that increase or decrease their risk is essential. Once identified, these factors can become targets of intervention or prevention to improve mental health among young men with poor PSC due to real or perceived adiposity.

Consequently, the current study focused on men dissatisfied with their adiposity by screening out underweight students as well as those whose overweight status was due to hypermuscularity. Thus, by design, all men with low PSC in our sample perceived themselves as having too much body fat. Creating a relatively homogeneous sample allowed for the investigation of the relationship between low PSC due to dissatisfaction with adiposity and symptoms of depression. Our aim was to test the ability of various facets of SC to act as protective factors and weaken the relationship between low PSC and depression. That is, we hypothesized that despite their being dissatisfied with their actual or perceived level of body fat, men who felt good about themselves as family members, social beings, or students would report fewer symptoms of depression than men with low PSC who did not have at least one area of functioning that allowed them to feel good about themselves. In addition, we assessed the relative contribution of different dimensions of SC (i.e., Physical, Family, Social, and Academic) to the prediction of depression in college men.

Method

Participants

Undergraduate men (N = 303) from a large public university in the south central United States completed the study. We discarded data from participants who spent too little time to have likely read the questionnaires carefully (n = 11). We used participants’ reported height and weight to calculate body mass index (BMI): BMI = (weight in kilograms)/(height in meters2) (Centers for Disease Control and Prevention, 2016). To control for differing psychosocial factors and, potentially, risk or protective factors experienced by underweight men, we screened participants with a BMI of less than 18.5 out of the analyses (n = 8). Also, we removed body‐dissatisfied men whose reported discontent was likely due to a desire for more muscularity (n = 10) or fitness (n = 7). Muscularity‐driven men were identified as men who reported weight lifting as their sport and engaged in more than the weekly physical activity guidelines (U.S. Department of Health and Human Services, 2008). Finally, because increased muscle mass is generally encouraged in the male collegiate athletic environment, participants on university or club (i.e., football, hockey, rugby) teams were also removed from the sample (n = 28). In summary, our intentionally more homogeneous sample of 239 men included college men with BMIs in the normal, overweight, or obese ranges who, if dissatisfied with their physical selves, were primarily concerned with the perception or reality of too much body fat. Notably, our homogeneous sample still included a majority (i.e., 60%) of participants from the random sample, as being or fearing overfat is normative among undergraduate men in the United States (Gillen & Lefkowitz, 2006).

Our sample was demographically representative of the university from which we drew it. Participants were predominantly White (69.5%), traditional age (i.e., 82% 18–24 years old), and heterosexual (91%). Latino (11.7%) and Black (9.6%) students also were represented in the sample. However, the 22 men from other ethnic minorities (e.g., Asian American, Native American, biethnic/biracial) were too few to expect them to represent their ethnic group as a whole accurately and were excluded in analyses of ethnic differences. A plurality of participants were in their first year of college (33%), and the rest were evenly distributed across classes.

Instruments

Demographic questionnaire. Participants spent 5 minutes answering demographic questions, including participation in sport and workout behaviors.

Self‐concept. The 82‐item Tennessee Self‐Concept Scale, Second Edition (TSCS:2; Fitts & Warren, 1996), aims to measure an individual's self‐reported sense of self on the basis of actions, preferences, and feelings. The TSCS:2 measures six dimensions of SC (i.e., Physical, Family, Social, Academic, Moral, and Personal). Moral SC was not addressed in the current study because of recognized problems with its validity (Marsh & Richards, 1988). Personal SC was not included in the current study because it approximates global self‐esteem, a construct that has been widely studied and consistently related to positive outcomes (e.g., Dishman et al., 2006; Orth, Robins, Trzesniewski, Maes, & Schmitt, 2009).

Items on the TSCS:2 are rated on a 5‐point Likert‐type scale ranging from 1 (always false) to 5 (always true; Fitts & Warren, 1996), with raw scores on each subscale converted to T scores. A T score ≥ 60 indicates that the individual tends to feel content, whereas a T score ≤ 40 on any subscale signals dissatisfaction with that particular facet of SC. Each subscale measures individuals’ perception of their competency, worth, and positive feelings about a specific facet of SC. PSC primarily assesses individuals’ satisfaction with their physique and weight, such as “I am neither too fat nor too thin.” The Family SC subscale measures feelings of value and relationship quality within individuals’ family system, such as “My family will always help me.” Social SC quantifies how individuals perceive their friendliness, social skills, and competence in developing relationships (e.g., “I get along well with other people”). The Academic SC subscale assesses individuals’ views of their competence in school settings and how they feel others view their capabilities in their academic environment. For example, one item states, “I will never be as smart as other people” (reverse scored; Fitts & Warren, 1996).

Validity reviews support both concurrent and construct validity of the four SC subscales used in the current study (Brown, 1998). With one exception, the internal consistency reliabilities of the TSCS:2 were at least acceptable with the current sample (α = .75–.81). The Academic SC scale initially had weak internal consistency (α = .68). Following best practices, we examined the effect of each item's removal on internal consistency and removed one item, “I am not as smart as the people around me,” to achieve acceptable reliability (i.e., α = .75; Hair, Anderson, Tatham, & Black, 1998). Prorated Academic SC scale scores allow for the use of normed T scores.

Depression. Participants reported their symptoms of depression and other types of psychological distress on the Symptom Checklist–90–Revised (Derogatis, 1994). Respondents considered how they felt and behaved over the past 7 days and chose from 0 (not at all) to 4 (extremely) in response to brief descriptions of 90 symptoms. Subscale scores are converted to T scores, with T ≥ 63 considered “clinically elevated.” An example of one of the 13 items assessing depression is “Feeling low in energy or slowed down.” The internal consistency reliability for Symptom Checklist–90–Revised depression in our sample was good (α = .89).

Procedure

We obtained approval to conduct this study through the university's institutional review board. Male undergraduate students volunteered through the psychology department's research portal to earn extra credit. Participants logged into the confidential and secure website to read the informed consent notice. Those who agreed to participate completed the three measures described earlier and were presented with a debriefing letter.

Results

We used SPSS, Version 22.0, and R (R Corp Team, 2013) to conduct the data analysis. For the moderation analyses, we used the PROCESS macro (Hayes & Preacher, 2014) within SPSS. Missing data were examined and addressed using multiple imputation practices (Schlomer, Bauman, & Card, 2010). All assumptions of multiple regressions were met.

We tested differences across demographic groups on key variables for potential confounds. The relatively positive PSC of African American men as compared with Latino and White men were noted; however, we did not include ethnicity as a covariate because of our desire to retain the 22 men from other ethnic minority groups in the sample. Correlational analyses revealed significant relationships (p < .001) between PSC, depressive symptoms, and the other three dimensions of SC, with absolute values of the Pearson product‐moment correlation coefficients ranging from .24 to .53. Higher SC in one facet was correlated with higher SC in other facets, as well as lower reported depressive symptoms.

We conducted three hierarchical multiple regressions (HMRs) to determine the main effects and interactions of Family, Academic, and Social SC on the relationship between PSC and symptoms of depression (see Table 1). We entered PSC into the model followed by a specific TSCS:2 subscale and the PSC × TSCS:2 subscale interaction term and then interpreted changes in R2 to test for moderation. We also used the software package R to calculate the relative importance of each dimension of SC in predicting depression.

Table 1

Family, Social, and Academic SC as Moderators of the Relationship Between Physical SC and Depression

Note. SC = self‐concept as assessed by Tennessee Self‐Concept Scale, Second Edition.

p < .05.

p < .01.

p < .001.

In each of the HMRs, the main effects of all four dimensions of SC were significant individual predictors of symptoms of depression (see Table 1). The Family SC model accounted for 21%, the Social SC model accounted for 22%, and the Academic SC model accounted for 23% of the variance in PSC. Moreover, the interactions of PSC with Family SC (R = .50, p < .001), as well as with Social SC (R = .48, p < .001), were also significant (see Figure 1 for Social SC results). The interaction of PSC and Academic SC was not significant (R = .49, p = .081).

Figure 1

Social Self‐Concept Moderation Model

Note. Symptoms of depression T score: M = 64.03, SD = 10.49. Physical self‐concept T score: M = 46.89, SD = 9.18.

Specifically, the results of the Family SC moderation revealed significant main effects of Family SC and PSC, as well as a significant interaction (see Table 1). As hypothesized, nearly all men reported more symptoms of depression if they had low Family SC than if they had high Family SC; this effect became significantly more pronounced as PSC dropped. The second HMR revealed significant PSC and Social SC main effects (see Table 1). As hypothesized, Social SC significantly moderated the relationship between PSC and depressive symptoms. Similar to the results for Family SC, among men with low PSC, participants reported significantly more symptoms of depression if they were also low in Social SC than if they had high Social SC (see Figure ).

The protective effects of average or higher Family and Social SC were very similar across models, in that our data indicated that men with the lowest Family SC and Social SC scores in the sample also reported the most symptoms of depression; however, among college men with the highest PSC, there appeared to be a paradoxical relationship such that those with high Family SC reported more symptoms of depression than those with low Family SC (n = 6). Nevertheless, having high PSC was so predictive of low depression that even men from that group with the lowest Family SC still scored below the clinical cutoff on the depression scale.

The third HMR with Academic SC revealed a main effect for PSC, and Academic SC was significant (see Table 1). However, the strong negative relationship between PSC and depression was not significantly altered by Academic SC. We conducted an analysis of variance with the Academic SC level as the independent variable and Symptom Checklist–90–Revised scores as the dependent variable to understand the relationship of these two variables further. Results revealed that men with high Academic SC reported significantly lower depressive symptoms than men with either low or average Academic SC, F(2, 62) = 9.42, p < .001.

The simple multiple regressions with PSC, Family SC, Social SC, and Academic SC as simultaneous predictors were significant, F(4, 234) = 19.51, p < .001, accounting for 25% of the variance in depressive symptoms. More specifically, lower Physical (b = –.24, p < .001), Social (b = –.21, p < .014), and Academic SC (b = –.21, p < .001) were significant unique predictors of depressive symptoms, but Family SC (b = –.06, p < .447) was not. Although the majority of the variance in depression scores was accounted for by the shared variance of the SC domains (i.e., 15.6%), squared semipartial correlations revealed that, respectively, Academic, Physical, and Social SC accounted for an additional 4%, 3%, and 2% of the variance in students’ depression scores. A statistical bootstrapping procedure using all possible orders of predictors (relimp in R package, Version 0.9‐6; Firth, 2006) allocates shared variance proportionately and calculates the relative importance of each SC predictor as a percentage of the accounted variance. In magnitude, these findings were 8.02% for Social SC, 7.76% for PSC, 6.62% for Academic SC, and 2.61% for Family SC.

Discussion

We examined how different aspects of SC—Family, Social, and Academic—influence the relationship between PSC and depressive symptoms in a sample of college men that was intentionally homogeneous concerning the type of body dissatisfaction. Whereas Academic SC was not a significant moderator, Family and Social SC were. Specifically, low PSC and symptoms of depression are significantly correlated, and feeling positive about one's family or social relationships appears to have a buffering effect on that. Despite poor PSC, relationally satisfied students report fewer symptoms of depression than their low PSC peers who do not have adequate Family or Social SC. Our results indicate that having low Family SC and low PSC has an additive effect on the likelihood of experiencing clinically significant depressive symptoms. Family SC significantly changed the relationship between PSC and depressive symptoms, but the protective function of good Family SC only operated for men with low PSC. Specifically, among men with below‐average PSC, high Family SC was associated with depression scores that were significantly lower than those of men with low Family SC.

Men who experience positive Family SC may have been raised in relatively more psychologically healthy environments than their peers with negative Family SC. A secure and loving family is an essential factor in helping children grow into happier and more resilient adults who have internalized a positive sense of self (e.g., Bowlby, 1973; Groh et al., 2014; Harris, 1999). Men from such backgrounds could struggle with their body weight or be dissatisfied with their appearance without it fueling the sense of worthlessness and pessimism that accompanies depression. Even when men with strong Family SC reported negative views of their own body, the emotional support, unconditional love, and sense of belonging that these men have may serve as protective factors. Despite a strong moderating effect, the unique contribution of Family SC is negligible. That is, Family SC shares much of its variance with other facets of SC, suggesting that helping men improve how they feel about themselves academically, socially, or physically could be good treatment goals for depressed college men.

For men who have negative perceptions of their bodies and appearance, positive views of one's adequacy in nonfamilial relationships (high Social SC) serve to buffer depressive symptoms. College men with positive Social SC may be more likely to engage in social activities and feel more connected to a local support system. Having close friends in college may offer not only emotional support but instrumental support as well (e.g., friends who wake you when you oversleep, share their class notes, or offer rides). At the same time, it is not difficult to imagine how negative evaluations of one's friendliness and social competence could contribute to depressive symptoms in the college, especially among body‐dissatisfied men.

The relative importance analyses indicate that Social SC and PSC both make significant contributions to the prediction of depressive scores. College campuses tend to be environments that demand social and physical competency, often offering a time period where students begin to examine and develop social identities (Karp, Holmstrom, & Gray, 1998) and physical habits (Leslie, Sparling, & Owen, 2001). Forging lifelong friendships and being very socially active are iconic aspects of the college experience. Our data support the notion that, for men, Social SC is strongly related to PSC and about equally important in predicting depression.

Low PSC may also be a function of exposure to media‐driven masculine ideals that become internalized and may fuel engaging in a comparison process that leads to unrealistic expectations, possibly influencing the development of depressive symptoms. Men who feel socially accepted by others, however, experience fewer depressive symptoms than men who lack such acceptance and feel attainment of a physical ideal is required for social approval. Despite negative images toward overweight people in the media, men who feel positively about their social relationships, even if they have low PSC, experience fewer depressive symptoms overall.

Although Academic SC did not affect the relationship between PSC and depression, Academic SC accounted for the largest percentage of unique variance in depression scores. That is, despite significant correlations among all five of our key variables, low sense of competence as a student is a unique and significant predictor of depression, even among college men who feel positive about other facets of their SC. Our post hoc analyses suggest that men with average or negative views of themselves as a student (i.e., Academic SC T < 55) experienced significantly more depressive symptoms than peers with high Academic SC. One explanation of the salience of Academic SC to mood is that the role of student is central to the identity of full‐time undergraduates; therefore, academic successes and failures may disproportionately affect their well‐being.

Counseling and Administrator Implications

For men with negative perceptions about their physical appearance who experience depressive symptoms, outreach programs and counselors can work to address and develop Social SC and PSC. In intervention planning for college men, enhancing clients’ views of their Social SC through social skills building and programs to assist them in making friends may help protect against or address current experiences of depression. Considering its high correlation with PSC, developing a more positive Social SC may help to build a positive view of one's physical self. Specific outreach programs include those that address social skills education, social‐event information, and activity fairs to increase student networks. Interventions may include interpersonal processing groups and social skill practice. Structured college environments that actively encourage engagement in clubs or activities may help increase social competency.

Interventions to boost Academic SC may consist of resources such as tutoring, supplemental instruction, recitation classes, and career counseling, despite one's PSC, which may help to address depressive symptoms. Also, administrators in higher education should consider failing grades or a significant drop in grade point average as more than just a need for academic intervention. The alert system that many schools have in place to identify students who are struggling academically could be expanded so that college and university counseling centers reach out to these students as well. Their programs should aim to bolster areas of SC and enhance activities that decrease risk for depression—for example exercise (Adame, Radell, Johnson, & Cole, 2003), good sleep hygiene (Allgöwer, Wardle, & Steptoe, 2001), and social support and stress management (Wang, Cai, Qian, & Peng, 2014). Moreover, the relationship between Academic SC and symptoms of depression is most likely bidirectional. Although struggling academically is a stressor that could contribute to negative feelings and poor sleep and even trigger depressive episodes, it is well‐established that symptoms of depression contribute to decreased academic performance and Academic SC (Deroma, Leach, & Leverett, 2009; Hysenbegasi, Hass, & Rowland, 2005). Within interpretations, it is important to remember that investigating men's dissatisfaction with their physical selves is much more complicated than investigating the same construct (i.e., desire is to be thinner) in women. Findings focus on the largest subsample of body‐dissatisfied men, those for whom PSC is due to the reality of having excess body fat. Our findings, therefore, should not be applied to men with low PSC who primarily want to be heavier, stronger, or more muscular.

Limitations and Future Directions

Several limitations require discussion and may prompt additional research. First, all participants were from one university, limiting the generalizability of findings. Information from a variety of universities (i.e., metropolitan, rural, private, liberal arts, elite) would contribute significantly to the generalizability. Also, participant responses were gathered in a cross‐sectional manner, leaving the directionality of all correlations unknown.

Furthermore, with respect to our negative findings regarding Academic SC as a moderator, Siemsen, Roth, and Oliveira (2010) demonstrated that common method variance could inflate or deflate multivariate regression models, making interactions more difficult to detect. African American men in the current study's sample had significantly higher PSC scores compared with White men, although depression scores did not differ. Our results may not apply to African American men in that results were interpreted on the basis of a majority White sample. More research is needed to understand how PSC and other facets of SC may be specifically protective (or not) concerning depression for men of different racial identities, sexual orientations, and weight. Furthermore, collecting longitudinal data on how changes in college semester grade point averages may relate to depressive symptoms would aid in the understanding of how Academic SC may be a risk or protective factor for college men.