The Role of Aspirations and Obligations in Explaining the Relationship between Identity Discrepancies and Psychological Distress

Abstract

Using a national sample of adults, we examine the relationship between identity discrepancies and mental health in spouse and worker identities. Building on previous research, we predict that discrepancies between how individuals want to be with respect to a particular identity (aspirations) and perceptions of how others view them in that identity (reflected appraisals) will be associated with depressive symptoms. Alternatively, discrepancies between how individuals feels they should be (obligations) and reflected appraisals will be associated with anxiety symptoms. We further examine whether identity salience moderates the relationship between discrepancies and distress. We find aspiration discrepancies are associated with depression as predicted in the spouse identity but not for the worker identity. With respect to obligation discrepancies, we find evidence for the predicted relationships for the spouse and worker identity only when identities are considered salient. We discuss the implications of our findings for the development of identity models of distress.

Keywords

aspirations obligations identity discrepancy distress

For over 20 years, sociological social psychologists have advanced models linking identities and psychological well-being (e.g., Burke 1991, 1996; Thoits 1991, 1999). According to Peggy Thoits (1999:345), it is “virtually impossible to develop a theory of the etiology of mental illness without thinking about self and identity issues.” Similar to biological and sociological models of stress, identity theorists conceptualize stressors in terms of external conditions, threats, or challenges to the system (e.g., Pearlin et al. 1981; Selye 1956; Wheaton 1991) and distress as an internal, subjective, or behavioral response to stressors (Burke 1991; Wheaton 1991). Also consistent with much of the mental health literature, identity theorists have focused on generalized distress as the likely outcome of identity-related stressors (Burke 1991, 1996; Thoits 1991, 1992). While symptoms of distress represent a common outcome of stressful circumstances, some mental health and identity scholars have emphasized the importance of examining different forms of distress or emotional outcomes to better understand the context in which stress occurs (e.g., Brown and Harris 1979; Burke and Stets 2009; Horwitz 2002; Stets and Burke 2000) .

In this article, we examine identity relevant stressors (discrepancies) and symptoms of two distress outcomes: depression and anxiety. To do so, we test a theoretical model that combines aspects of perceptual control processes¹ proposed originally by Peter Burke (1991, 1996), as well as psychological processes proposed by E. Tory Higgins (1987, 1989), to account for discrepancies that occur within and across role identities. In particular, identity discrepancy theory (IDT) expands the concept of the identity standard to include self-meanings based on aspirations and obligations associated with identities (Large and Marcussen 2000; Marcussen 2006; Marcussen and Large 2003). By examining these specific aspects of identity, we may be able to more precisely predict the types of distress individuals experience as a result of discrepancies that occur between their self-view and their perception of how they are viewed by others.

While initial tests of IDT have appeared in previous research (Marcussen 2006; Marcussen and Large 2003), these studies relied on nonprobability samples of students, resulting in an examination of transitional identities. The current study is unique in that it uses a national probability sample to examine aspects of identity for two primary role identities: spouse and worker. Our measurement strategy is informed by a pilot study aimed at developing measures that better capture distinguishing aspects of aspirations and obligations within these identities. Finally, we examine the influence of identity salience in the context of aspiration-based and obligation-based discrepancies. While these widely held work and family-based identities are likely to have some degree of shared expectations, they may also vary to some extent in terms of their individual meaning. Although our theoretical arguments do not suggest differential predictions for these roles with respect to the general identity process, we expect that specific meanings (i.e., aspirations and obligations) and relevance (i.e., salience) associated with these identities may vary by identity type. Including both identities will provide a more robust test of identity theory and provide an opportunity to generalize findings more broadly to adult social roles.

Theoretical Models of Self, Identity, and Distress

Perceptual control theory (Burke 1991, 1996; Burke and Stets 2009) offers a formal model of identity that predicts distress resulting from the inability to verify one’s sense of self. The model specifies four basic components of an identity: the identity standard, self-relevant perceptions (or input), a comparator, and output. The identity standard is the content or meaning the identity holds for the individual. Self-relevant perceptions are the individual’s understandings of situational elements that inform the individual about the self and constitute the input into the identity system. They include both an individual’s perceptions of the situation and the individual’s perception of how others view the situation, or reflected appraisals (Riley and Burke 1995). The comparator is the process by which the congruence between the identity standard and the input is assessed (Burke 1991, 1996). The output is “meaningful behavior” oriented toward maintaining consistency between self-relevant perceptions and the identity standard. The identity process involves a continuous assessment of the consistency between the identity standard (what our social roles mean to us) and reflected appraisals (how we believe others see us in our identities). Individuals are motivated to achieve a match between the standard for an identity and reflected appraisals in the situation. When there is incongruence between the identity standard and reflected appraisals, the individual acts in ways to alter that situation and thereby change the feedback from the social situation to be more consistent with the identity standard. In general, the identity process operates smoothly, making minor adjustments as necessary. When the process is disrupted, distress occurs. Persistent discrepancy (or nonverification) of identities has been linked to generalized distress (e.g., Burke and Harrod 2005; Cast and Burke 2002) and a range of emotional responses (e.g., Stets 2003; Stets and Asencio 2008; Stets and Burke 2014a; Trettevik and Grindal 2016).

Perceptual processes within identity theory provide a useful model for explaining distress, but until recently (see Burke and Stets 2009), the theory did not provide a means to distinguish outcomes. Self-discrepancy theory (Higgins 1987, 1989) focuses on such distinctions, particularly with respect to two key mental health outcomes: depression and anxiety. According to Higgins, the self consists of three domains: actual, ideal, and ought. The actual self refers to a representation of the attributes that the individual (or some other) believes he or she possesses. The ideal self refers to a representation of the attributes that the individual (or some other) wishes or hopes to possess. The ought self refers to a representation of the attributes that the individual (or some other) believes the individual should possess. In essence, the ideal self reflects aspirations for the individual and the ought self reflects feelings of obligation.

Similar to perceptual control theory, Higgins (1987, 1989) described a process whereby discrepancies between the self-concept (actual self) and self-guides (ideal and ought selves) result in distress. However, self-discrepancy theory also specifies the type of emotional reaction that will result from discrepancies. According to Higgins, and consistent with previous work linking cognitive processes to specific emotional outcomes (e.g., Lazarus 1968; Stein and Jewett 1986), discrepancies that occur between the actual self and the ideal self represent the failure to achieve a goal, and therefore produce dejection-related outcomes (e.g., disappointment, sadness). Alternatively, discrepancies that occur between the ought and actual self represent a fear that something negative will happen as a consequence, and therefore produce agitation-related outcomes (e.g., fear, restlessness). The links between self-guides and distress have been demonstrated in Higgins’s early studies (e.g., Higgins 1987, 1989) and recently extended to address the mental health consequences of discrepancies that occur within specific components of one’s self, such as ideals related to body image (e.g., Halliwell and Dittmar 2006; Vartanian 2012).

Identity-discrepancy Theory: Integrating Two Models of Distress

IDT (Large and Marcussen 2000; Marcussen 2006; Marcussen and Large 2003) combines aspects of these psychological and sociological models of distress to predict depression and anxiety that result from inconsistency within identities. Whereas identity theory has called attention to the importance of meaning to the study of social roles, IDT incorporates aspects of self-discrepancy theory to differentiate self-meanings into aspirations and obligations. Rather than conceptualize ideals and oughts in the context of a global self, identity theories conceptualize the self as comprised of multiple identities. As such, IDT combines aspects of self-discrepancy theory and perceptual control systems to suggest that different types of discrepancies can occur within the various social roles we hold, as these roles likely vary in meaning.

For IDT, the identity standard in Burke’s model is expanded to contain identity meanings associated with aspirations and obligations, which are in turn related to specific distress outcomes. For example, whereas traditional perceptual control models might look at a general set of characteristics of a “spouse” identity (e.g., I am attentive as a spouse) as comprising the identity standard, IDT argues that a person may have a “spouse” identity containing both aspirations (e.g., “I want to be forgiving as a spouse”) and obligations (e.g., “I should be patient as a spouse”). Although these aspects of the identity standard may overlap to some extent, there are also times where they may differ. Such a conceptualization of an identity allows for more specific predictions about the composition of distress. This elaboration is a key advantage to IDT, as it increases the utility of discrepancy theories within sociological social psychology, and at the same time provides a formal model of stress and mental health that emphasizes role meaning as a key factor in understanding psychological outcomes.

Mental health scholars agree that there is a need for theoretical models that can discriminate between anxiety and depression (Aldao et al. 2010; Mennin et al. 2008; Watson and Kendall 1989). While often highly associated, the two disorders are physiologically and psychologically distinct (Beck 1967; Finlay-Jones and Brown 1981). Focusing on social status differences in mental health, stress researchers have also pointed to the importance of determining pathways to distinct disorders that might better help us understand the distribution, causes, and social consequences of distress (e.g., Brown and Harris 1979; Rosenfield, Lennon, and White 2005). Along those lines, identity theorists have identified such differences to be rooted in variation in the meanings that social roles hold (e.g., Burke 1991; Large and Marcussen 2000; Simon 1992, 1997; Thoits 1991). Moreover, extensions of identity models of emotion (Burke and Stets 2009) advance an argument for further specifying predictions of emotional outcomes based on the source of meaning and the source of discrepancy (self or other). More recently, identity scholars have examined the role that “gaps” between ought/actual dimensions and ideal/actual dimensions play in explaining the relationship between identity discrepancy and positive and negative emotional outcomes in the student identity (Trettevik and Grindal 2016). We similarly seek to make finer distinctions about the relationships between social roles and distress.

The Role of Identity Salience in the Relationship between Discrepancies and Distress

In addition to specifying the relationship between discrepancies and distress, identity theorists have also acknowledged aspects of the identity process that may serve to moderate or condition this relationship. According to Burke (1991), the more organized and well established an identity, the more sensitive it will be to inconsistent feedback. Related to this, the more salient (Burke 1991; Large and Marcussen 2000) or the more accessible (Higgins 1987, 1989) an identity, the more likely the discrepancy will be associated with distress. Although specified as a potential contributing factor in the identity process, few studies have directly examined salience in discrepancy models. Studies that have considered salience in identity models of distress more broadly have produced mixed results (e.g., Marcussen, Ritter, and Safron 2004; Thoits 1991, 1992). We speculate that this may be due to differences in how salience is defined and measured.

While conceptual definitions and measurement of identity salience have varied to some degree in the literature (see Thoits 2013), salience generally refers to the relative placement of an identity in a broader set of identities, and the extent to which that placement influences the activation of the identity in a social context. Salience has been measured in terms of relative rankings, as well as overall ratings of how important or likely to be mentioned an identity might be (Thoits 2013). We define salience in terms of the likelihood that an identity will be invoked across social situations (e.g., Stryker and Serpe 1982; Stryker and Statham 1985). This conceptualization is also consistent with Higgins’s (1987) notion of accessibility. Here, salience does not refer to the activation of an identity itself, but to the probability that the identity will be activated (Stets and Burke 2000).

According to the Stryker and colleagues, the more salient the role, the more likely it is that one will invest in behaviors associated with that role (Stryker and Serpe 1994; Stryker and Statham 1985). Similarly for discrepancy theories, the more salient the identity, the more likely one is to respond to discrepancies within the identity in terms of emotional and behavioral responses (Burke 1991; Large and Marcussen 2000). While the spouse and worker identities are generally considered important and culturally valued, the extent to which they are salient (or likely to be invoked) may vary across individuals. We argue that such variation is likely to condition the relationship between discrepancies and mental health outcomes.

Hypotheses

As articulated in previous work (e.g., Burke 1991,1996; Burke and Stets 2009; Marcussen 2006; Marcussen and Large 2003), we expect that identity discrepancies will be related to distress. Drawing from assumptions of IDT, we further predict specific forms of distress will be associated with different types of identity discrepancy. Specifically, we extend the identity standard to include aspirations and obligations. In the context of IDT, discrepancies between aspirations and reflected appraisals are primarily associated with depressive symptoms, whereas discrepancies between reflected appraisals and the obligation dimension of the identity standard are primarily associated with anxiety symptoms. These assumptions form the basis of our first set of hypotheses:

Hypothesis 1 (H1): Aspiration discrepancies will be positively related to depressive symptoms.

Hypothesis 1a (H1a): Aspiration discrepancies will be more strongly associated with symptoms of depression than with symptoms of anxiety.

Hypothesis 2 (H2): Obligation discrepancies will be positively related to anxiety symptoms.

Hypothesis 2a (H2a): Obligation discrepancies will be more strongly associated with symptoms of anxiety than with symptoms of depression.

As discussed above, the relationship between identity discrepancies and distress may be conditioned by factors such as how likely an individual is to invoke the identity across social situations. Consistent with Burke (1991), we propose that discrepancies that occur in more salient identities will have a more intense effect on mental health outcomes. Our second set of analyses tests that assumption in the context of IDT:

Hypothesis 3 (H3): The relationship between aspiration discrepancies and depression will be stronger at higher levels of identity salience.

Hypothesis 4 (H4): The relationship between obligation discrepancies and anxiety will be stronger at higher levels of identity salience.

Method

Data and Sample

The data for this research come from a national telephone survey of 1,409 adults. The survey was designed to examine aspects of identity theory and psychological well-being. The sample was drawn using a multistage cluster sample of telephone numbers (both landline and cell) in the continental United States. The survey included five identities. All respondents were questioned about two of those identities that were randomly assigned from the total number of identities for which they were screened as eligible. We focus here on the spousal and work identities only.

The data were collected between September and December of 2011, with a response rate of 47 percent. This response rate is consistent or better than similar samples using random digit dialing (RDD).² While research on response rates shows that the effects of nonresponse are not necessarily threatening to the accuracy or generalizability of survey data (see Keeter et al. 2006), it is important to note that our sample differs from population characteristics in some regards. Comparisons with census data indicate that our sample has more respondents who are women, white, college-educated, and older relative to the general population. To some extent, these differences may reflect the fact that our sample consists of individuals who hold at least two adult social roles. Given this variation in demographic characteristics, results should be interpreted with some caution.

Measures

Dependent variables

Our study focuses on two measures of psychological distress: depressive and anxiety symptoms, both of which are measured using the Hopkins Symptom Checklist–25 (HSCL-25, Derogatis et al. 1974; Hesbacher et al. 1980). The HSCL-25 was adapted from the Symptom Checklist–90–revised (SCL-90-R; Derogatis 1977) and includes items that measure depression as well as subjective and somatic indicators of anxiety. This measure is ideal for our purposes because it is easy to administer and has been suggested to have clearly separate factors for anxiety and depression (Morgan, Wiederman, and Magnus 1998). The HSCL-25 also has high discriminant validity and reliability when used to measure distress in both clinical and community settings.³

The following script is used for the HSCL-25: “Now I am going to read from a list of symptoms or problems that people sometimes have. As I read these items, decide how much the symptoms bothered or distressed you in the last week, including today.” Responses range from one (not at all) to four (extremely), and higher scores indicate more depressive or anxiety symptoms, respectively. The depression and anxiety symptom variables used in the analyses are comprised of the factor scores for depressive and anxiety symptoms. As shown in Table 1, factor loadings are moderate to high for all items, and both measures are highly reliable.

Table 1.

Factor Loadings for Depression and Anxiety Scale Items (N = 1,409).

	Factor loadings
Depressive symptoms (15 items)
Feeling low in energy, slowed down	.588
Poor appetite	.583
Feeling everything is an effort	.702
Loss of sexual interest or pleasure	.504
Feeling blue	.773
Difficulty falling asleep or staying asleep	.477
Worrying too much about things	.591
Crying easily	.603
Feeling hopeless about the future	.706
Feelings of worthlessness	.778
Blaming oneself for things	.625
Thoughts of ending one’s life	.585
Feeling lonely	.678
Feeling trapped or caught	.714
Feeling no interest in things	.678
Eigenvalue	6.774
Reliability (α)	.910
Anxiety symptoms (10 items)
Being suddenly scared for no apparent reason	.761
Feeling fearful	.695
Nervousness or shakiness inside	.752
Spells of terror or panic	.800
Heart pounding or racing	.681
Feeling restless, not being able to sit still	.563
Headaches	.448
Feeling faintness, dizziness, or weakness	.676
Trembling	.729
Feeling tense or keyed up	.551
Eigenvalue	5.060
Reliability (α)	.885

Note. Factor analyses conducted separately for depressive and anxiety symptoms using maximum likelihood extraction and mean replacement for missing values.

Independent variables

The independent variables in our analyses are aspiration and obligation identity discrepancies in two work and family role domains: spouse and worker. Screening questions determined whether individuals were currently married and/or employed. For the purposes of this study, we focus only on individuals who are actively engaged in one or both of these identities.

Discrepancies are calculated using respondents’ ratings of adjectives that represent characteristics or attributes of the various identities. Specifically, for each identity, respondents rated seven adjectives on a scale from 0 (not at all) to 10 (extremely) in terms of how much they (1) want to be, (2) feel obligated to be, (3) think others see them as, and (4) see themselves. For example, for the identity SPOUSE and the adjective LOYAL, respondents would be asked to what extent, as a spouse, they wanted to be loyal, felt obligated to be loyal, thought others viewed them as loyal, and viewed themselves as loyal. These four questions capture aspirations, obligations, reflected appraisals, and self-views, respectively. Our focus here is to determine whether aspirations and obligations (rather than a more general measure of self-view) might help distinguish distress outcomes. Therefore, for the purposes of this study, we focus on these two dimensions of the identity standard relative to reflected appraisals.⁴

The adjectives we use for the above scales were chosen to reflect aspects of each identity based on the results of a pilot study. We began with a list of 200 adjectives that were deemed “highly meaningful” to respondents in that they resonated most with respondents and had a high degree of shared meaning (Anderson 1968). Using this list, we enlisted independent raters to identify a list of 20 adjectives that corresponded with each of our identities. From those lists, we selected 10 adjectives to include in the pilot study. Using the pilot study data, we then conducted factor analysis with maximum likelihood estimation and further reduced the number of adjectives to be included in the present study to seven per identity. As shown in Table 2, all but one of the adjective factor loadings are moderate to high for both aspirations and obligations in the two identities in this study, and all measures are highly reliable. The one exception is the adjective “independent,” which does not load well with the other spouse aspiration items. However, we retain it in the analyses so that the overall number of adjectives is consistent across identities, and the aspiration and obligation adjectives are the same within each respective identity. The reliability of this scale remains high (α = .71).

Table 2.

Factor Loadings for Aspiration and Obligation Adjective Scale Items by Identity.

	Spouse (N = 409)			Worker (N = 417)
		Factor loadings			Factor loadings
	Adjectives	Aspirations	Obligations	Adjectives	Aspirations	Obligations
	Forgiving	.629	.802	Loyal	.568	.646
	Attentive	.603	.795	Enthusiastic	.678	.787
	Patient	.696	.840	Efficient	.574	.690
	Cooperative	.714	.867	Cooperative	.403	.696
	Interesting	.498	.629	Ambitious	.610	.734
	Independent	.063	.441	Conscientious	.512	.653
	Reasonable	.728	.781	Agreeable	.579	.660
Eigenvalue		3.101	4.329		2.903	3.903
Reliability (α)		.709	.893		.740	.866

Note. Factor analyses conducted separately for aspirations and obligations for each identity using maximum likelihood extraction and mean replacement for missing values.

For each identity, we calculate squared and linear discrepancy scores, consistent with previous research (Stets and Burke 2014a). We calculate identity discrepancies using the following steps: (1) sum adjective ratings for aspirations, obligations, and reflected appraisals, respectively; (2) divide each scale by the total number of items; (3) compute discrepancies by subtracting the aspiration scale from the reflected appraisal scale and the obligation scale from the reflected appraisal scale, respectively; and (4) square discrepancies.

Moderating variables

In addition to testing the basic assumptions of the IDT model, we examine the moderating role of salience in the relationship between discrepancy and distress. For each identity, respondents were asked four questions about how certain they are that they would tell a person whom they were meeting for the first time (i.e., a coworker, friend of a close friend, friend of a family member, person of the opposite sex) about that identity on a scale from one (almost certainly would not) to four (almost certainly would). The salience items for each identity were summed and divided by the total number of items. Higher scores indicate higher salience for a given identity.

Control variables

In the analyses presented here, we control for sex (female = 1), age, employment status (1 = employed), racial/ethnic background (1 = nonwhite), marital status (1 = married), education (1 = less than high school, 2 = high school graduate, 3 = some college, 4 = college graduate, 5 = postgraduate), and having children living at home (1 = children living at home). Marital status is not included in the spouse model, and employment status is not included in the worker model. Tables 3 and 4 show descriptive statistics for the variables included in the analysis as well as the bivariate correlations between the key theoretical variables in this study.

Table 3.

Descriptive Statistics for All Study Variables.

	N	M/proportion	SD	Range
Demographic variables
Age	1,409	55.28	14.810	18 to 99
Female	1,409	0.65		0 to 1
Married	1,409	0.64		0 to 1
Education (categories)^a	1,401	3.39	1.130	1 to 5
Employed	1,409	0.46		0 to 1
Nonwhite	1,396	0.15		0 to 1
Children living at home^b	1,409	0.34		0 to 1
Dependent and moderating variables
Anxiety symptoms (factor scores)	1,409	0.00	0.959	−0.89 to 4.37
Depressive symptoms (factor scores)	1,409	0.00	0.950	−0.64 to 4.69
Spouse salience	392	2.78	0.790	1 to 4
Worker salience	398	2.87	0.599	1 to 4
Independent variables (discrepancies)
Spouse identity
Linear spouse aspirations	379	−1.01	0.968	−4.29 to 1.29
Squared spouse aspirations	379	1.96	2.746	0.00 to 18.37
Linear spouse obligations	379	0.65	2.022	−3.43 to 9.57
Squared spouse obligations	379	4.50	11.413	0.00 to 91.61
Worker identity
Linear worker aspirations	402	−0.58	0.749	−3.57 to 4.86
Squared worker aspirations	402	0.89	1.798	0.00 to 23.59
Linear worker obligations	402	0.49	1.570	−3.57 to 8.86
Squared worker obligations	402	2.70	7.438	0.00 to 78.45

Note. Among those with children (N = 1,123), the proportion with children living at home is 0.42.

Education: 1 = less than high school; 2 = high school graduate; 3 = some college; 4 = college degree; 5 = postgraduate.

The proportion of the total sample (N = 1,409) with children (living at home or not) is 0.80.

Table 4.

Bivariate Correlations for Aspiration and Obligation Discrepancy and Distress Measures.

	1	2	3	4	5	6	7	8	9	10
1. Linear spouse aspiration discrepancy	1.000
2. Squared spouse aspiration discrepancy	−.818	1.000
3. Linear spouse obligation discrepancy	.274	−.211	1.000
4. Squared spouse obligation discrepancy	n.s.	n.s.	.784	1.000
5. Linear worker aspiration discrepancy	.550	−.192	.202	n.s.	1.000
6. Squared worker aspiration discrepancy	n.s.	.429	n.s.	n.s.	−.370	1.000
7. Linear worker obligation discrepancy	.220	−.135	.623	.562	.267	−.134	1.000
8. Squared worker obligation discrepancy	n.s.	n.s.	.485	.684	n.s.	n.s.	.756	1.000
9. Anxiety symptoms	n.s.	.082	−.112	n.s.	n.s.	n.s.	n.s.	n.s.	1.000
10. Depressive symptoms	−.133	.169	−.170	−.097	n.s.	n.s.	n.s.	n.s.	.808	1.000

Note. All correlations are statistically significant at p ≤ .05 (two-tailed) or better. Ns vary by cell. Salience variables (not shown) are only significantly associated with one another (r = .348).

Analytic Strategy

We conduct two sets of structural equation models (SEMs) for each identity: spouse and worker. The first set of models tests the basic theoretical predictions of IDT, whereas the second set of models examines the moderating role of salience in the relationship between identity discrepancy and psychological distress. In all models, variables are calculated outside of SEM (as described above) and entered in SEMs as observed variables. We use this strategy to preserve degrees of freedom in our relatively small identity subsamples. Although latent variable analysis is beneficial in that it somewhat accounts for measurement error, the scales used in this study have been demonstrated to be reliable and valid, and thus, it is appropriate to include them as observed variables.⁵

We use structural equation modeling because it is ideal for simultaneously estimating relationships between multiple independent and outcome variables and it provides the information necessary to evaluate how well a model fits the pattern of relationships in the data. In all SEMs, we use full information maximum likelihood estimation. Following Burke and colleagues (Burke and Harrod 2005; Stets and Burke 2014a, 2014b), we examine the impact of squared discrepancies in models that control for linear discrepancies.⁶ Squared discrepancies are included because, like perceptual control models referenced above, we assume that distress increases as discrepancies (in either direction). As noted in previous work (e.g., Burke and Harrod 2005; Stets and Burke 2014a), if squared discrepancies significantly predict distress in models that control for linear discrepancies, then the self-verification (or consistency) assumption that discrepancies in any direction are distressing is supported. For the linear term, scores above zero indicate that individuals perceive that others see them more positively than they feel they should be or aspire to be (overevaluation), whereas scores below zero would indicate that individuals perceive that others see them less positively than they feel they should be or aspire to be (underevaluation).

In the moderation models, cross-product terms for discrepancy and salience are calculated and entered into the models as observed variables. To minimize the potential for multicollinearity, we standardize the discrepancy and salience measures before computing each interaction term and enter interaction terms for aspiration and obligation discrepancy (both linear and squared terms) in separate models for each respective identity.⁷ The salience models include all elements included in the base models, but add salience and discrepancy by salience interaction term variables. Causal paths between salience and the outcomes and interaction terms and the outcomes are estimated as well as correlations between all exogenous variables.

Results

Table 5 shows standardized coefficients for all of the key theoretical paths in the base IDT models and salience models for each of the two identities and for the relationships between control variables and outcome variables. This table does not report the coefficients for the relationships between control variables and identity discrepancies or the correlations between depressive and anxiety symptoms. In the SEMs presented in Table 5, correlations between depressive and anxiety symptoms are spouse identity, r = .776, p < .001, and worker identity, r = .781, p < .001. Despite being relatively high, these correlations indicate that depressive and anxiety symptoms are not equivalent. Fit statistics indicate that both base models fit the data very well, and that the salience models fit the data well. Although the chi-square statistics are significant for the salience models and the root mean square error of approximation (RMSEA) statistic for the worker salience model is slightly higher than the typical threshold considered to indicate a good fit, the chi-square to degrees of freedom ratio is 3:1 or smaller for both salience models, indicating an acceptable fit. All fit statistics are presented in the footnotes of Table 5. Below, we discuss the results of the SEMs for each set of hypotheses, focusing on the theoretical variables only.

Table 5.

Standardized Coefficients (Direct Effects) for Structural Equation Models Estimated Separately for Each Identity.

	Spouse base^a (N = 409)		Spouse salience^b (N = 409)		Worker base^c (N = 417)		Worker salience^d (N = 417)
	Depression	Anxiety	Depression	Anxiety	Depression	Anxiety	Depression	Anxiety
Age	.001	−.039	.001	−.048	.001	.001	.001	.001
Female	−.006	.001	−.003	.007	.001	−.022	−.006	−.029
Married	—	—	—	—	−.155**	−.106*	−.143**	−.093
Education	−.051	−.127*	−.055	−.132**	−.060	−.075	−.068	−.083
Employed	.028	−.025	.033	−.021	—	—	—	—
Nonwhite	.015	.007	.012	.003	.187***	.148**	.184***	.146**
Children at home	.053	.001	.056	.001	.009	.037	.014	.044
Linear aspiration discrepancy	.182	−.026	.177	−.034	.043	.033	.043	.035
Squared aspiration discrepancy	.268*	−.054	.270*	−.053	−.026	−.018	−.031	−.022
Linear obligation discrepancy	−.041	−.030	−.033	−.016	−.214**	−.062	−.223**	−.075
Squared obligation discrepancy	−.073	−.024	−.091	−.059	.135	.035	.176*	.084
Salience	—	—	−.066	−.084	—	—	.071	.074
Linear obligation discrepancy × Salience	—	—	−.018	−.002	—	—	−.007	−.026
Squared obligation discrepancy × Salience	—	—	.091	.153**	—	—	−.112	.132*

Note. Italicized parameters were constrained to a value of .001. CFI = comparative fit index; RMSEA = root mean square error of approximation.

χ² = 1.30, 4 df, p = .861, CFI = 1.000, RMSEA = .000; in addition to the constraints shown, employment status on linear and squared aspiration discrepancy was also constrained to a value of .001.

χ² = 31.51, 16 df, p = .012, CFI = .989, RMSEA = .049; in addition to the constraints shown, employment status on linear and squared aspiration discrepancy was also constrained to a value of .001.

χ² = 7.52,4 df, p = .111, CFI = .996, RMSEA = .046; in addition to the constraints shown, education on linear aspiration discrepancy and age on squared obligation discrepancy were also constrained to a value of .001.

χ² = 49.36,16 df, p = .000, CFI = .965, RMSEA = .049; in addition to the constraints shown, education on linear aspiration discrepancy and age on squared obligation discrepancy were also constrained to a value of .001.

p ≤ .05. **p ≤ .01. ***p ≤ .001 (two-tailed).

Aspiration Discrepancy Hypotheses

H1 and H1a pertain to the relationship between aspiration discrepancies and depression. Consistent with H1, we find that squared aspiration discrepancy is positively associated with depressive symptoms for the spouse (B = .268, p < .05). We also find support for H1a for the spouse identity; squared aspiration discrepancy is not significantly associated with anxiety symptoms, and Wald tests confirm that it is not statistically different from zero (p = .636) and not equal to the relationship between squared spouse aspirations and depressive symptoms (p = .000). We find no support for aspiration hypotheses in the worker identity.

Obligation Discrepancy Hypotheses

We also examine the relationship between obligation discrepancies and anxiety for each of the two identities. Squared obligation discrepancy is not associated with anxiety symptoms in either of the two identities, and thus, we do not find support for H2 or H2a in the base model. Interestingly, as shown in Table 5, linear obligation discrepancy in the worker identity is negatively associated with depressive symptoms (B = −.214, p < .01), indicating that underevaluation increases depressive symptoms and overevaluation decreases depressive symptoms in this identity. This pattern of results runs counter to our predictions.

Salience Hypotheses

Our final set of hypotheses examines whether salience moderates the relationship between aspiration discrepancies (H3), obligation discrepancies (H4), and distress. In analyses not shown, we do not find support for H3 in that the aspiration discrepancy-by-salience interactions were not statistically significant for the identities examined here. As shown in Table 5, we do find support for H4 in that salience moderates the relationship between squared obligation discrepancy and anxiety symptoms for both the spouse (B = .153, p < .01) and worker (B = .132, p < .05) identities.

Because our models test interaction terms that involve quadratic variables, the significant interaction for the focal variable (Squared Discrepancy × Salience) tells us that the form of the relationship between discrepancies and distress is altered by salience (Dawson 2014). In other words, the relationship between obligation discrepancy and anxiety symptoms is different for spouses and workers who report high compared with low levels of salience. To further examine the specific nature of these significant interaction effects for respondents with high and low salience in the spouse and worker identities, we plot them using coefficients for the main effects of linear and squared obligation discrepancy and salience on anxiety symptoms, coefficients for interactions between linear and squared obligation discrepancy terms and salience, and the means and standard deviations of these variables. These plots can be seen in Figures 1 and 2.

Figure 1.

Squared Spouse Obligation Discrepancy × Salience interaction.

Figure 2.

Squared Worker Obligation Discrepancy × Salience interaction.

For both identities, low salience is defined as one standard deviation below the mean and high salience is defined as one standard deviation above the mean. It should be noted that mean salience is just below three (on a scale with a maximum value of four) for both the spouse and worker identities. Thus, “low” and “high” salience should not be interpreted as such in an absolute sense but rather in reference to mean levels of salience in each identity. The x-axis of the plot for each identity is centered about the mean for linear obligation discrepancy in that identity and depicts one standard deviation below and above that mean on the left and right sides of the x-axis, respectively. The y-axis of both plots represents predicted values of anxiety symptoms and ranges from zero to four.

Beginning with the spouse identity plot (Figure 1), we see that the relationship between obligation discrepancy and anxiety takes on a curved shape at high and low salience and is roughly centered about zero on the x-axis. The curve for high salience is somewhat steeper than the curve for low salience, suggesting that H4, which predicted that the relationship between obligation discrepancy and anxiety would be stronger at higher levels of salience, is supported. The worker identity shows a similar pattern (Figure 2). The relationship between obligation discrepancy and anxiety takes on a curved shape at high and low salience and is roughly centered about zero on the x-axis. Again the curve for high salience appears somewhat steeper than the curve for low salience, though the difference between the two is slightly less pronounced. Thus, we find support for H4 in the worker identity.

Discussion

Building on psychological and sociological models of self and identity (e.g., Burke and Stets 2009; Higgins 1987), IDT posits that when identity discrepancies persist, individuals will experience increases in symptoms of depression and anxiety (Large and Marcussen 2000; Marcussen 2006; Marcussen and Large 2003). The theory further suggests that symptoms of depression are associated with one’s desires or aspirations within an identity, and symptoms of anxiety are associated with one’s sense of obligation within an identity. Consistent with perceptual control models and mental health research more generally, we suggest these patterns will be stronger at higher levels of salience. We test these ideas for the first time with a national sample of adults in two commonly held identities: spouse and worker.

There are two patterns that emerge in our findings that are particularly worthy of discussion. First, we find that while aspiration discrepancies operate as expected for the spousal identity, our expectations are not supported for the worker identity. Specifically, for individuals who report on the spousal identity, experiencing a discrepancy between how they feel they want to be in a given identity and how they perceive others view them in that identity is associated with higher levels of depressive symptoms. For the worker identity, however, aspiration discrepancies are not associated with distress. Accounting for the salience of spouse and worker identities did not change the relationship between aspiration discrepancies and distress.

Findings for obligation discrepancies show a different pattern, and one that is more dependent on the salience of identities. We do not find squared obligation discrepancies to be related to symptoms of anxiety in either the spouse or worker identity. Indeed, the only significant relationship we find with respect to obligation discrepancies is the significant and negative relationship between linear obligation discrepancy and depression for the worker identity only. This pattern is contrary to our expectations, but somewhat consistent with previous studies (Marcussen 2006; Marcussen and Large 2003) that find obligation discrepancies to behave somewhat differently than aspiration discrepancies. Importantly, when salience is considered, we see a pattern that is consistent with IDT predictions. For both the spouse and worker identities, the more salient the identity, the more that discrepancies between obligations and reflected appraisals are associated with anxiety (but not depressive) symptoms.

While our findings are mixed, we believe our study contributes to identity theory and the mental health literature in a number of ways. First, we extend an identity model of distress that emphasizes role meaning. Meanings attached to social roles have been a central concern for identity and mental health research, particularly in work that links the two (e.g., Marcussen et al. 2004; Simon 1997; Thoits 1991). According to Simon (1997), early theoretical arguments that focused on salience hierarchies alone failed to adequately assess concrete meanings that individuals assign to their roles. In her qualitative work, she found that individuals emphasized fulfilling personal goals and meeting responsibilities as defining features of what it means to be a spouse and worker (Simon 1997). Over the past two decades, perceptual control models have refined measures of identity based on dimensions that individuals believe represent “myself as” a spouse (e.g., Burke and Harrod 2005), parent (e.g., Tsushima and Burke 1999), and worker (e.g., Stets and Asencio 2008). IDT extends the conceptualization and measurement of meaning in perceptual control models by incorporating two dimensions of identity meaning that move beyond how the individual sees himself or herself in a role to incorporate desires (“I want to be”) and responsibilities (“I should be”) within these roles. A limitation of previous research was that the empirical distinctions between these two dimensions of the identity standard were somewhat unclear (Marcussen and Large 2003). Our measurement strategy for distinguishing between obligations and aspirations in this study shows a clearer pattern of difference both within and across identities. Extending Higgins (1987), and consistent with Simon (1997), we show that these dimensions of identity are an important part of how individuals define their work and family roles, and should be considered in work examining the causes and consequences of stress in these identities.

Second, our study links these dimensions of the identity standard to specific mental health outcomes. Mental health scholars have emphasized the need to elaborate mechanisms that link social roles to distress and broaden the scope of mental health outcomes to better understand the link between social factors and variation in mental illness. IDT highlights these goals specifically. Focusing on squared discrepancies as prescribed by identity theory, we find that where aspirations are related to mental health, they are associated with depression. Obligations, on the contrary, are associated with anxiety when the worker and spousal identities were considered more salient. It is worth noting that we find the strongest support for the theory when the covariance between aspirations and obligations are at their lowest (spouse identity), and more mixed findings when the correlations between these two components of the identity standard are higher (worker identity). In other words, when distinctions between these two aspects of identity are clearer, depressive symptoms are more strongly associated with aspirations than obligations, and obligations are more strongly associated with symptoms of anxiety than are aspirations. Studies that examine only depression or anxiety, or that combine the two in a measure of generalized distress, may be underestimating the relationship between identities and mental health in general, as well as the relationship between role meaning and distress more specifically.

Generally speaking, failure to confirm one’s identity should be more distressing when the discrepancy occurs in identities that are highly salient compared with those that are less salient (Burke 1991, 1996). Nonetheless, salience has been overlooked or seemingly inconsequential in research examining identity and distress. Here we find that salience, defined as the probability of enacting an identity (Stryker and Serpe 1982), is instrumental in terms of understanding the relationship between how people feel they “should” be and how they believe others view them in the spouse and worker identities. Specifically, we see that the patterns for obligation discrepancies reflect a U-shaped relationship between discrepancy and anxiety when salience is taken into account. These findings bring to mind recent work that shows that when meanings in a situation are relevant, consistency effects emerge (Stets and Burke 2014a). Given the differences in the relative impact of salience in our models, they also point to potential differences in how salience interacts with specific identity meanings; whereby the relevance of the identity might be more (as in the case of obligations) or less (as in the case of aspirations) important in the identity process.

Limitations and Future Directions

This research provides a step toward extending an identity theory of distress; however, there are some limitations to consider when interpreting the findings and considering future directions. As noted earlier, an advantage of this study is that it draws on a national probability sample of adults. Previous studies examining identity processes tend to use student and/or convenience samples. While these studies have provided insight into important theoretical processes, they provide somewhat limited information about adult social roles. The mental health consequences of the spouse and worker roles have been studied broadly in the mental heath literature. Our data allow us to better understand identity processes in the context of these two key roles. That said, we recognize that there are limitations to our sample. As discussed above, our response rate, while considered quite good for an RDD survey, may raise concerns regarding nonrandom differences in our sample. Some of these differences likely reflect the fact that respondents hold one or both of these adult roles. It is further possible, however, that older and more highly educated respondents are more established in these roles, which might result in differences in reported identity meanings and/or identity salience. Women and white respondents may also respond differently to questions regarding identity meaning and distress relative to their counterparts. While theoretically the identity process should operate similarly across groups, and we take steps to control for these demographic factors in our analyses, differences between the sample and population should be considered when generalizing the results.

Second, while including salience as a key component of the discrepancy process is a contribution of this study, there are some potential limitations to using these standard measures of salience that should be addressed in future research. One issue worthy of consideration is whether there is a relationship between the concealable nature of an identity and the salience of that identity. For instance, are identities that are more easily detected (e.g., the spousal identity) more or less likely to encourage discussion or behavior when meeting someone for the first time relative to identities that are less easily known? A second concern is the lack of variability in salience for highly normative and valued roles such as these. Indeed, in our sample we find that the spousal and worker identity are very similar in terms of their average salience. Moving forward, the inclusion of additional factors such as identity commitment and prominence may be important. Researchers have argued that the ability to exit an identity, the amount of time spent in an identity, and the relative importance of the identity for sense of self are all likely to increase the effects of disruption in that identity on well-being (Burke and Harrod 2005; Oatley and Bolton 1985; Swann, Pelham, and Krull 1989). Extending that logic, it is also possible that these additional pieces of information about identity meaning might be more or less influential depending on whether one is considering their aspirations or obligations within a given identity.

Looking at the two identities we examined here, we can begin to see how such factors might help to clarify existing patterns. Focusing on the worker identity, for example, it is reasonable to assume that individuals might respond differently when speaking about a transitional or part-time job than they would a professional career or longtime occupation. It is quite possible that aspirations regarding the worker identity are related to the extent to which the individual’s current job is reflective of their overall sense of self, where obligations may reflect more specific expectations about having a job and/or being a good worker. In addition, though the roles that we investigate might both be considered obligatory in nature (e.g., Thoits 2003), the costs associated with exiting work and family roles may differ. While the inclusion of salience provided insight into the relationship between obligation discrepancies and anxiety for the spouse and worker identities, it is possible that other factors might be more influential with respect to the relationship between aspirations and depression. Future research should consider whether identity commitment and prominence add more information about the structural connections, normative expectations, and self-relevance associated with these identities in ways that shed light on differences we found between these two identities, as well as underscore relationships that are more consistent with our theoretical predictions.

Related to this point, another important step in this work will be to examine multiple identities simultaneously (Burke and Stets 2009). Our study highlights a comparison of two identities, but we do not examine the relationship between the identities, nor do we examine them in the context of other identities that individuals might hold. Consistent with identity theory, we argue that the process we describe occurs for each identity a person holds, but it likely does so within the context of multiple and potentially simultaneously activated identities. When the meanings (including aspirations and obligations) across identities are shared, the verification process for one identity assists with the verification of the others (Deaux 1992). Shared meanings across identities should also increase the likelihood that they are activated together, leading to similar levels of salience and commitment (Stets and Burke 2003). However, when identities have different or competing meanings, the verification process may be altered or hindered (Burke and Stets 2009). To the extent that aspirations and/or obligations conflict or overlap across and within identities, we might expect the relationship between discrepancy and distress to become more complicated. These ideas are also in line with research that emphasizes the function of self-complexity and role meaning in explaining multiple role occupancy and mental health (e.g., Bolger et al. 1989; Linville 1985; Simon 1995).

Finally, we believe this research could be developed further by considering a broader array of psychological and emotional outcomes. While research shows that depression and anxiety can be distinguished in terms of their causes and consequences, there is considerable comorbidity in these measures of distress. Future research should explore specific types of depressive and anxiety symptoms, as symptoms within these two forms of distress may vary quite a bit. For instance, measures of social anxiety that are distinct from other types of anxiety tend to also be more distinct from depressive symptoms (Higgins 1987). This work should also explore other aspects of emotion that have been clearly linked to aspiration and obligation discrepancies in self-discrepancy research and in more recent models that incorporate oughts and ideals into perceptual control models (e.g., Trettevik and Grindal 2016), including sadness and dejection (aspiration-related) and guilt and fear (obligation-related). These emotional outcomes may be more sensitive to different aspects of identity meaning and therefore particularly appropriate for understanding well-being in the context of identity discrepancy.

Conclusion

This research takes a second look at a theory that has previously been tested with convenience samples of student populations. Using national data and examining adult social roles, this study presents a step toward understanding the relationship between aspiration and obligation discrepancies and psychological distress in more stable role identities. Our findings suggest the importance of examining variation in identity meaning in an effort to better specify pathways that link social roles and mental health.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for research, authorship, and/or publication of this article: This research was supported by a National Science Foundation grant awarded to the first author (Award: 1024342).

Notes

Author Biographies

Kristen Marcussen is an associate professor of sociology at Kent State University. Her research focuses on explaining variation in mental health as a function of social roles and identity processes. Most recently, she is using identity models to examine stigma and self-processes among individuals with mental illness. Her work has been funded by the National Science Foundation and appears in outlets such as Social Psychology Quarterly, Journal of Health and Social Behavior, and Sociological Perspectives.

Mary Gallagher is an assistant professor of sociology at Kent State University at Stark. Her research interests include self, identity, mental health, and the connections between them. Her current research is situated within identity theory and utilizes survey data to examine identity discrepancy processes, identity commitment and centrality, and reflected appraisals in the context of adult social roles.

References

Aldao

Amelia.

Mennin

Doug S.

Linardatos

Eftihia

Fresco

David M.

2010. “Differential Patterns of Physical Symptoms and Subjective Processes in Generalized Anxiety Disorder and Unipolar Depression.” Journal of Anxiety Disorders 24:250–59.

American Association for Public Opinion Research. 2011. Standard Definitions: Final Dispositions of Case Codes and Outcome Rates for Surveys. 7th ed.

American Association for Public Opinion Research

Anderson

Norman H.

1968. “Likableness Ratings of 555 Personality-trait Words.” Journal of Personality and Social Psychology 9:272–79.

Beck

Aaron T.

1967. Depression: Clinical, Experimental, and Theoretical Aspect. New York: Hoeber.

Bolger

Niall

DeLongis

Anita

Kessler

Ronald

Wethington

Elaine

. 1989. “The Contagion of Stress across Multiple Roles.” Journal of Marriage and Family 51:175–83.

Brown

Tirril

Harris

George W.

1979. Social Origins of Depression. A Study of Psychiatric Disorder in Women. New York: Tavistock Publications.

Burke

Peter J.

1991. “Identity Processes and Social Stress.” American Sociological Review 56: 836–49.

Burke

Peter J.

1996. “Social Identities and Psychological Stress.” Pp. 141–74 in Psychosocial Stress: Perspectives on Structure, Theory, Life Course, and Methods, edited by Kaplan

Howard B.

Orlando, FL: Academic Press.

Burke

Peter J.

Harrod

Michael M.

2005. “Too Much of a Good Thing?” Social Psychology Quarterly 68:359–74.

10.

Burke

Peter J.

Stets

Jan E.

2009. Identity Theory. New York: Oxford University Press.

11.

Cast

Alicia D.

Burke

Peter J.

2002. “A Theory of Self-esteem.” Social Forces 80:1041–68.

12.

Dawson

Jeremy F.

2014. “Moderation in Management Research: What, Why, When, and How.” Journal of Business and Psychology 29:1–19.

13.

Deaux

Kay

. 1992. “Personalizing Identity and Socializing Self.” Pp. 9–33 in Social Psychology of Identity and the Self-concept, edited by Breakwell

London, England: Academic Press.

14.

Derogatis

Leonard R.

1977. The SCL-90 Manual I: Scoring Administration and Procedures for the SCL-90. Baltimore, MD: Clinical Psychometric Research.

15.

Derogatis

Leonard R.

Lipman

Ronald S.

Rickels

Karl

Uhlenhuth

E. H.

Covi

Lino

. 1974. “The Hopkins Symptom Checklist (HSCL): A Self-report Symptom Inventory.” Behavior Sciences 19:1–15.

16.

Finlay-Jones

Robert

Brown

George W.

1981. “Types of Stressful Life Events and Onset of Anxiety and Depressive Disorders.” Psychological Medicine 11:803–15.

17.

Halliwell

Emma

Dittmar

Helga

. 2006. “Associations between Appearance-related Self-discrepancies and Young Women’s and Men’s Affect, Body Satisfaction, and Emotional Eating: A Comparison of Fixed Item and Participant-generated Self-discrepancies.” Personality and Social Psychology Bulletin 32:447–58.

18.

Hesbacher

Peter T.

Rickels

Karl

Morris

Rebecca J.

Newman

Rosenfeld

1980. “Psychiatric Illness in Family Practice.” Journal of Clinical Psychiatry 41:6–10.

19.

Higgins

E. Tory

. 1987. “Self-discrepancy: A Theory Relating Self and Affect.” Psychological Review 94:319–40.

20.

Higgins

E. Tory

. 1989. “Self-discrepancy Theory: What Patterns of Self-beliefs Cause People to Suffer?” Advances in Experimental Social Psychology 22:93–136.

21.

Horwitz

Allan V.

2002. “Outcomes in the Sociology of Mental Health and Illness: Where Have We Been and Where Are We Going?” Journal of Health and Social Behavior 43:143–51.

22.

Keeter

Scott

Kennedy

Courtney

Dimock

Michael

Best

Jonathan

Craighill

Peyton

. 2006. “Gauging the Impact of Growing Nonresponse on Estimates from a national RDD telephone survey.” Public Opinion Quarterly 70:759–79.

23.

Large

Michael D.

Marcussen

Kristen

. 2000. “Extending Identity Theory to Predict Forms and Degrees of Psychological Distress.” Social Psychological Quarterly 63:49–59.

24.

Lazarus

Arnold A.

1968. “Learning Theory and the Treatment of Depression.” Behavior Research and Therapy 6:83–88.

25.

Linville

Patricia W.

1985. “Self-complexity and Affective Extremity: Don’t Put All Your Eggs in One Cognitive Basket.” Social Cognition 3:94–120.

26.

Marcussen

Kristen

. 2006. “Identities, Self-esteem, and Psychological Distress: An Application of Identity-discrepancy Theory.” Sociological Perspectives 49:1–24.

27.

Marcussen

Kristen

Large

Michael D.

2003. “Using Identity Discrepancy Theory to Predict Psychological Distress.” Pp. 151–66 in Advances in Identity Theory and Research, edited by Burke

Peter J.

Owens

Timothy J.

Serpe

Richard T.

Thoits

Peggy A.

New York: Kluwer Academic/Plenum Press.

28.

Marcussen

Kristen

Ritter

Christian

Safron

Deborah

. 2004. “The Role of Identity Salience and Commitment in the Stress Process.” Sociological Perspectives 47:289–312.

29.

Mennin

Douglas S.

Heimberg

Richard G.

Fresco

David M.

Ritter

Michael R.

2008. “Is Generalized Anxiety Disorder an Anxiety or Mood Disorder? Considering Multiple Factors as We Ponder the Fate of GAD.” Depression and Anxiety 25:289–99.

30.

Morgan

C. Don

Wiederman

Michael W.

Magnus

Ryan D.

1998. “Discriminant Validity of the SCL-90 Dimensions of Anxiety and Depression.” Assessment 5:197–201.

31.

Oatley

Keith

Bolton

Winifred

. 1985. “A Social-cognitive Theory of Depression in Reaction to Life Events.” Psychological Review 92(3):372–88.

32.

Pearlin

Leonard I.

Menaghan

Elizabeth

Lieberman

Morton A.

Mullan

Joseph T.

1981. “The Stress Process.” Journal of Health and Social Behavior 22:337–56.

33.

Riley

Anna

Burke

Peter J.

1995. “Identities and Self-verification in the Small Group.” Social Psychology Quarterly 58:61–73.

34.

Rosenfield

Sarah

Lennon

Mary Clare

White

Helene

. 2005. “Mental Health and the Self: Self-salience and the Emergence of Internalizing and Externalizing Problems.” Journal of Health and Social Behavior 46:326–40.

35.

Selye

Hans

. 1956. The Stress of Life. New York: McGraw-Hill.

36.

Simon

Robin W.

1992. “Parental Role Strains, Salience of Parental Identity, and Gender Differences in Psychological Distress.” Journal of Health and Social Behavior 33:25–35.

37.

Simon

Robin W.

1995. “Gender, Multiple Roles, Role Meaning, and Mental Health.” Journal of Health and Social Behavior 36:182–94.

38.

Simon

Robin W.

1997. “The Meanings Individuals Attach to Role-Identities and their Implications for Mental Health.” Journal of Health and Social Behavior 38:256-274.

39.

Stein

Nancy L.

Jewett

Janet L.

1986. “A Conceptual Analysis of the Meaning of’ Negative Emotions: Implications for a Theory of Development.” Pp. 410–43 in Measurement of Emotion in Infants and Children, edited by Izard

Caroll E.

Read

Peter B.

New York: Cambridge University Press.

40.

Stets

Jan E.

2003. “Justice, Emotion, and Identity Theory.” Pp. 105–22 in Advances in Identity Theory and Research, edited by Burke

Peter J.

Owens

Timothy J.

Serpe

Thoits

Peggy A.

New York: Kluwer Academic.

41.

Stets

Jan E.

Asencio

Emily K.

2008. “Consistency and Enhancement Processes in Understanding Emotions.” Social Forces 86:1055–78.

42.

Stets

Jan E.

Burke

Peter J.

2000. “Identity Theory and Social Identity Theory.” Social Psychology Quarterly 63:224–37.

43.

Stets

Jan E.

Burke

Peter J.

2003. “A Sociological Approach to Self and Identity.” Pp. 128–52 in Handbook of Self and Identity, edited by Leary

Mark

Tangney

June

. New York: Guilford Press.

44.

Stets

Jan E.

Burke

Peter J.

2014a. “Emotions and Identity Nonverification.” Social Psychology Quarterly 77:387–410.

45.

Stets

Jan E.

Burke

Peter J.

2014b. “Self-esteem and Identities.” Sociological Perspectives 57:409–33.

46.

Stets

Jan E.

Serpe

Richard T.

2013. “Identity Theory.” Pp. 331–60 in Handbook of Social Psychology, 2nd ed., edited by DeLamater

John

Ward

Amanda

. New York: Springer.

47.

Stryker

Sheldon

Serpe

Richard T.

1982. “Commitment, Identity Salience, and Role Behavior: A Theory and Research Example.” Pp. 199–218 in Personality, Roles and Social Behavior, edited by Ickes

Knowles

E. S.

New York: Springer-Verlag.

48.

Stryker

Sheldon

Serpe

Richard T.

1994. “Identity Salience and Psychological Centrality: Equivalent, Overlapping, or Complementary Concepts?” Social Psychology Quarterly 57:16–35.

49.

Stryker

Sheldon

Statham

Anne

. 1985. “Symbolic Interaction and Role Theory.” Pp. 311–78 in The Handbook of Social Psychology, edited by Lindzey

Gardner

Aronson

Elliot

. New York: Random House.

50.

Swann

William B.

Jr. Pelham

Brett W.

Krull

Douglas S.

1989. “Agreeable Fancy or Disagreeable Truth? Reconciling Self-enhancement and Self-verification.” Journal of Personality and Social Psychology 57:782–91.

51.

Thoits

Peggy A.

1991. “On Merging Identity Theory and Stress Research.” Social Psychology Quarterly 54:101–12.

52.

Thoits

Peggy A.

1992. “Identity Structures and Psychological Well-being: Gender and Marital Status Comparisons.” Social Psychology Quarterly 55:236–56.

53.

Thoits

Peggy A.

1999. “Self, Identity, Stress, and Mental Health.” Pp. 345–68 in Handbook of the Sociology of Mental Health, edited by Aneshensel

Carol S.

Phelan

Jo C.

New York: Kluwer Academic/Plenum Press.

54.

Thoits

Peggy A.

2003. “Personal Agency in the Accumulation of Multiple Role-identities.” Pp. 179–94. In Advances in Identity Theory and Research, edited by Burke

P. J.

Owens

T. J.

Serpe

R. T.

Thoits

P. A.

New York: Kluwer Academic/Plenum Press.

55.

Thoits

Peggy A.

2013. “Volunteer Identity Salience, Role Enactment, and Well-being: Comparisons of Three Salience Constructs.” Social Psychology Quarterly 76:373–98.

56.

Trettevik

Ryan

Grindal

Matthew

. 2016. “The Ideal and Ought Self-guides and the Affective Consequences of Identity Verification.” Pp. 601–26 in New Directions in Identity Theory and Research, edited by Stets

Jan E.

Serpe

Richard T.

New York: Oxford University Press.

57.

Tsushima

Teresa

Burke

Peter J.

1999. “Levels, Agency, and Control in the Parent Identity.” Social Psychology Quarterly 62:173–89.

58.

Vartanian

Lenny R. (2012).

“Self-discrepancy theory and body image.” Pp. 711-717 in Encyclopedia of Body Image and Human Appearance, edited by Cash

Thomas F.

San Diego, CA: Academic Press.

59.

Watson

David

Kendall

Phillip C.

1989. “Understanding Anxiety and Depression: Their Relation to Negative and Positive Affective States.” Pp. 3–26 in Anxiety and Depression: Distinctive and Overlapping Features, edited by Kendall

Phillip C.

Watson

David

. San Diego, CA: Academic Press.

60.

Wheaton

Blair

. 1991. “Sampling the Stress Universe.” Pp. 77–114 in Stress and Mental Health: Contemporary Issues and Prospects for the Future, edited by Avison

William R.

Gotlib

Ian H.

New York: Plenum Press.