Identity Nonverification,Emotions,and Cognitive and Behavioral Responses

Emotions

Recently, it has been argued that the emotions that emerge from the identity verification process carry meaning and that these meanings and situational identity meanings will affect cognitive and behavioral responses to identity nonverification (Burke and Stets 2023). These correspond to Points 1 and 2, respectively, in Figure 1. For example, people may feel anger following identity nonverification, perhaps because they think that others are responsible for them not being verified. This is an external attribution in which they deny responsibility for the identity nonverification (Stets and Burke 2005). Anger is a negative emotion often directed at others who are perceived to be the cause of a blocked goal (Berkowitz and Harmon-Jones 2004). Here, the goal is to achieve verification, which has been hindered. Cognitively, individuals may think others are at fault for the nonverification, and behaviorally, they may express their negative feelings and ask others to apologize.

In this research, we begin to examine the theoretical idea that emotions resulting from identity nonverification should be similar in meaning to the cognitive and behavioral responses driven by those emotions. The meanings should correspond in valence and in content. In terms of valence, positive emotions should lead to positive thoughts and behaviors, and negative emotions should produce negative thoughts and behavior. In terms of content, when a person reports an emotion, the meanings of that emotion should correspond to the cognitive and behavior responses often tied to that emotion. Given these observations, anger should be associated with unhelpful, negative thoughts, such as blaming others, and negative behaviors, such as expressing negative feelings.

We examine the emotions and resulting cognitive and behavioral responses during the COVID-19 pandemic. During COVID-19, individuals experienced disruption in their work life, home life, friendships, and intimate relationships. Consequently, people likely had trouble playing out and verifying identities across multiple domains. Figure 2 presents our heuristic model.

OPEN IN VIEWER

Figure 2.

Heuristic Model of Identity Nonverification, Emotions, and Coping Strategies

Identity theorists have focused largely on general positive and negative emotions to the exclusion of an array of specific emotions that emerge from the nonverification process (Burke and Stets 2023; Stets and Trettevik 2014). Although earlier theoretical work suggested how the attribution process might influence specific emotions in response to identity nonverification, with internal attributions producing feelings such as sadness and shame and external attributions fostering emotions such as hostility and anger (Stets and Burke 2005), it did not discuss how emotions might relate to the cognitive and behavioral responses to those emotions, which is the focus of this research. We expand our understanding of the role of specific emotions in identity theory by studying 10 positive and negative emotions that may relate to the nonverification process and, in turn, the choice of cognitive and behavioral coping strategies. Paths 1 and 2 in Figure 1 of the identity model are of primary interest to us.

We examine five primary emotions (anger, disgust, fear, happiness, and sadness) and five emotions understood as secondary emotions (guilt, hope, pride, shame, and sympathy) (Turner and Stets 2005). Primary emotions are thought to be the core emotions from which many other emotions are derived. They are believed to be hardwired in the human neuroanatomy because of their fitness-enhancing value. Secondary emotions are more socially constructed and may emerge from experiencing one or more of the primary emotions. For example, hope is a combination of happiness and fear (Kemper 1987; Turner 2000).

Coping Strategies

“Coping” refers to the strategies that people use to prevent being harmed by aspects of social experiences that disrupt them from carrying out their roles in social life and that adversely penetrate their emotional lives (Pearlin et al. 1981; Pearlin and Schooler 1978). It is people’s response to stressful situations (Lazarus and Folkman 1984). Identity nonverification is a situation of stress because the continuous operating process of enacting and verifying one’s identity has been interrupted (Burke 1991).

Although the identity verification process involves people maintaining perceived identity-relevant meanings in the situation to be consistent with the meanings in their identity standards, this process also involves the thoughts and behaviors that people undertake to modify the situation and the meanings contained therein to prevent a problem from occurring in the first place. Acting to keep identity-relevant situational meanings in accord with identity standard meanings prevents feeling distress. When situational meanings stray from the identity standard and identity nonverification has emerged, that control has been diminished or lost. This brings about the autonomic response felt as distress that grabs our attention so we can act more deliberately to restore situational meanings to those contained in the identity standard. This is where coping strategies become relevant.

Effective coping strategies change meanings in the situation in particular directions and in particular amounts to make situational meanings match identity standard meanings. The thoughts and behaviors that will do that can only be judged by the perceiver and need to be constantly monitored and controlled. We explore how people’s feelings are associated with their coping strategies. Following others (Carver, Scheier, and Weintraub 1989), coping is a multidimensional construct. In this research, we rely on a commonly used measure of coping, the Brief COPE scale (Carver 1997), which captures the strategies people tend to use to cope, including addressing the stressor, seeking social support, and avoiding the stressor (Farley et al. 2005; Kato 2015; Shamblaw, Rumas, and Best 2021; Solberg, Gridley, and Peters 2022).

Following previous work, we use the labels most commonly applied to the three strategies as “active,” “support,” and “disengaged” coping in the Brief COPE scale (Solberg et al. 2022). Active coping aims to address the stressor. Support involves reaching out to others or one’s religion to manage stress. Disengaged coping entails avoiding the stressor or distancing oneself from it.

During the COVID-19 pandemic, people tended to use active and support styles of coping more than disengaged coping (Afifi and Afifi 2021; Logel, Oreopoulos, and Petronijevic 2021; Park et al. 2020; Polizzi, Lynn, and Perry 2020). Individuals tried to seek solutions to their stress, remaining hopeful and positively reframing their situation, keeping busy, and seeking support from others (Kar, Kar, and Kar 2021; Logel et al. 2021; Shamblaw et al. 2021). Although different aspects of the pandemic were associated with the coping strategies we examine, we focus on the role of identity nonverification and the different emotions respondents experienced.

Our main hypotheses are with the effects of identity nonverification on negative emotions, positive emotions, and coping. Because we do not have strong theory for relating specific emotions to specific cognitive and behavioral responses, our analysis for this part of the study is exploratory. Thus, we do not offer hypotheses regarding how each of our 10 emotions relate to the different coping strategies we examine. Our hypotheses are broad in order to cover our fundamental expectations. We anticipate the following:

We try to understand which emotions bring about which coping responses to better understand the emotions-coping relationship in the identity nonverification process. Because we focus on identity nonverification arising from the effects of the COVID-19 pandemic, we include identity nonverification that occurred prior to the pandemic as a control. We also include demographic controls, discussed in the following, to rule out spurious connections between identity nonverification and our endogenous variables.

Sample

An online survey was conducted by Qualtrics to obtain a nonprobability quota sample of 840 respondents in the spring of 2021, a year into the pandemic. The goal was to obtain an approximately equal number of respondents across different social groups, including gender, race/ethnicity, age, and income. We were not interested in trying to generalize our results to a population, but rather, we wanted to assess the robustness of the emotion-coping relationship across important social categories. We examined four identities that were selected by most respondents as being the most difficult to play out during the pandemic: friend identity, romantic identity, worker identity, and family identity. Our final sample is 620 individuals. These are the respondents from our four racial/ethnic groups who selected one of the four identities. Omitted are those who were not in our four racial/ethnic groups and who did not identify one of the four identities as the most difficult for them to enact during the pandemic.

The 620 respondents were distributed as follows: (1) 49% men and 51% women; (2) 19% Asian, 21% Black, 24% Hispanic, and 36% non-Hispanic White; (3) 41% low income (<$50,000), 32% middle income ($50,000–$100,000), and 27% high income (>$100,000); and (4) 37% age 18–34, 38% age 35–54, and 25% 55 and older. The identities were 32% for the friend identity, 20% for the romantic identity, 25% for the worker identity, and 23% for the family identity. We examined the variability of the estimated effects of a structural equation model across the two gender categories, the four racial/ethnic categories, and the four identities.

Measures

To measure the identity processes resulting from nonverification, we first ascertained for respondents the identity they had the most difficulty enacting during the pandemic. We asked respondents to think about the various ways they might see themselves, and we gave them the examples of seeing themselves as a student, friend, democrat, and an American. Then, they were given a list of nine areas in which people might have difficulty playing out who they are. They were asked to select the one area they had the most difficulty playing out during the past year of the pandemic, including their occupation, friendships, romantic relationships, political party, family, gender, race or ethnicity, nationality, and religion. Each respondent selected the single identity they had the most difficulty playing out during the pandemic. A tenth item indicated they had “no problems” in any of the areas listed.

We limited our analyses to the areas people most often cited for which they had frequent problems playing out who they were. Areas with fewer than 50 respondents were eliminated. This left the areas of respondents’ occupation, friendships, romantic relationships, and family. We labeled these the worker identity, friend identity, romantic identity, and family identity, respectively. These four identities represented 78% of the selected identities chosen by respondents. Each identity was coded 1 if the respondent identified difficulty playing it out; otherwise, it was coded 0.

Identity nonverification was measured by asking respondents how much difficulty they had in playing out their chosen identity during the pandemic. Response categories ranged from “not at all difficult” to “extremely difficult” (coded 0–10). Higher scores reflected greater identity difficulty or nonverification. This measure was based on self-appraisals, or an individuals’ own assessments as to how well they were meeting the identity standard they set for themselves (Burke and Stets 2023). Although earlier research often has focused on reflected appraisals (how much individuals think others see them in the same way they see themselves), relying on a reflected appraisal measure is limiting when people are better able than others to judge how they are doing in a situation in terms of playing out their identity. Thus, we measured verification from respondents’ direct assessment of themselves. This measure has been used in other research (Burke and Harrod 2021; Grindal, Kushida, and Nieri 2021).

Ten specific emotions that might influence behavioral choices were measured, including anger, disgust, fear, guilt, happiness, hope, pride, sadness, shame, and sympathy. The questions on emotions occurred after respondents answered questions about their identity. They were asked to “Please indicate the degree to which you have felt [emotion] over the past 2 weeks, with zero being ‘Not at All’ and six being ‘Intensely.’” ¹ Four emotions had a positive valence (happiness, hope, pride, and sympathy), and the other six had a negative valence (anger, disgust, fear, sadness, guilt, and shame).

During analysis of the emotions, we discovered that respondents varied in their use of the emotion scales, with some using more extreme values in their reporting of emotions than others. This resulted in an artifactual confounding of “emotional volatility” with the level of specific emotions that were felt. To counteract this, the variable emotionality was created by first standardizing the measure of each of the 10 emotions and then calculating the average level of each respondent’s standardized responses across all 10 emotions. Those who routinely had higher levels of response across the emotions (positive and negative) thus had higher scores on the emotionality variable. By including emotionality as a control variable in the analysis, the level of any emotion reported by a respondent was, in effect, measured from their average level of reported emotion, thus removing the confounding effect of emotional volatility.

For coping strategies used during COVID-19, respondents indicated how much, during the pandemic, they had been doing each of 28 coping tactics in the Brief COPE Scale (Carver 1997). The activities covered 14 coping strategies with two items for each strategy. Response categories for the items included “I haven’t been doing this at all,” “I’ve been doing this a little bit,” “I’ve been doing this a medium amount,” and “I’ve been doing this a lot” (coded 1–4, respectively).

Studies have found that the 28 items factor into the three general coping strategies (Solberg et al. 2022), which we also found. Our factor analysis of the responses to the 28 items, their loadings, and the reliability of each of the three scales are in Appendix A. The 28 individual items clustered into three strong principal component factors, which we labeled “active,” “support,” and “disengaged” (Solberg et al. 2022). Each of the factors had a high omega reliability (Heise and Bohrnstedt 1970). To measure use of each coping strategy, the responses on the items for each dimension were averaged, with a higher score representing more frequent use of the coping strategy. We also analyzed the 28 separate strategies to understand whether respondents’ emotions were associated with these strategies.

Across the 28 strategies, we see both cognitive and behavioral approaches. Active coping involved 10 responses that addressed the stressor (see Appendix A), and there were about an equal number of cognitive and behavioral techniques. Cognitive tactics included, for example, “thinking hard about what steps to take” (Item 5) and “trying to see it in a different light, to make it seem more positive” (Item 10). Behavioral approaches included, for instance, “doing something to think about it less, such as going to the movies, watching TV, reading, daydreaming, sleeping, or shopping” (Item 6) and “turning to work or other activities to take my mind off things” (Item 8).

Support coping contained six responses (again, see Appendix A), with a preponderance of them behavioral responses, such as “praying or meditating” (Item 2) and “getting emotional support from others” (Item 6). Finally, disengaged coping involved 12 responses that appeared to distance individuals from the source of the stress, and we found a mixture of cognitive and behavioral responses. Cognitive responses included, for example, “giving up the attempt to cope” (Item 4) and “refusing to believe that it has happened” (Item 8). Behavioral responses involved, for instance, “using alcohol or other drugs to help me get through it” (Item 2) and “saying things to let my unpleasant feeling escape” (Item 11).

We controlled whether respondents experienced nonverification of their selected identity prior to the pandemic (prior nonverification) to examine how nonverification is related to emotions and coping behaviors during the pandemic after removing the effects of prior nonverification. Respondents were asked whether they had difficulty with the identity they selected prior to the pandemic. Response categories were “no” or “yes” (coded 0/1, respectively). If respondents reported prior difficulty with their selected identity, prior identity difficulty was coded 1; otherwise, it was coded 0. Like the nonverification measure, this measure was based on respondents’ own assessment (self-appraisal) of how well they were meeting their identity standard.

We investigated whether the effects of nonverification on specific emotions and particular coping responses were consistent across separate groups, including gender, race/ethnicity, and the four identities. For gender, female was coded 0 for male and 1 for female. For race/ethnicity, respondents were shown seven racial/ethnic categories (and an “other” category) and asked to select their racial category, including all that applied. The categories were (1) American Indian or Alaska Native; (2) Asian; (3) Black or African American; (4) Hispanic, Latinx, Spanish; (5) Middle Eastern or North African; 6) Native Hawaiian or Other Pacific Islander; (7) White; or (8) some other race. Those respondents who self-selected into one of four racial/ethnic categories—Asian, Black, Hispanic, or White—fit the sampling frame. Respondents not selecting one of these categories were dropped. We coded these as three binary variables: Asian, Black, and Hispanic people (all coded 0 for no and 1 for yes). White people were the omitted category.

We also controlled for other factors, including age, income, and marital status, that may be related to our specific emotions or the coping strategies. For example, the more resources and the higher people’s position in the social structure along such dimensions as age and income, and the more social support they receive in social relationships such as marriage, the more they will experience positive emotions than negative emotions (Myers and Diener 2018; Schnittker 2008; Yang 2008). Similarly, a higher social structural position, for example, a higher income, should be associated with active coping compared to disengaged coping, and being married should be related to supportive coping than disengaged coping.

Age was measured in years. Income was measured by respondents selecting which of 16 categories contained their family income, from under $5,000 to $150,000 or more. Each person’s income was scored at the midpoint of the category in thousands of dollars. Married was coded 1 for currently married and 0 otherwise. Never married was coded 1 for those who had never been married and 0 otherwise. Separated/divorced/widowed was coded 1 for those who were separated, divorced, or widowed and 0 otherwise. Married was the omitted category.

Analysis

The model in Figure 2 was estimated using structural equations. In addition to the two nonverification variables (current and prior) and the emotionality variable, there were several exogenous background variables that were associated with the 10 emotions and the three coping strategies: active, support, and disengaged. The 10 emotions also were related to the coping strategies. The error terms on each of the 10 emotions were allowed to be correlated, as were the error terms on the three coping strategies. A second structural equation analysis was conducted using the 28 specific coping strategies as outcomes.

In each case, the analysis proceeded in two stages. In the initial stage, before the structural equation analysis, stepwise regressions were used to choose the emotions for the final model that best predicted the three coping strategies in Figure 2 after all the other variables were included. Specifically, in the first stage, before the emotions were potentially added in the stepwise procedure, we included the emotionality variable, the nonverification measures, and the controls. In this way, the final model for the second stage using structural equations included only those emotions that made an additional unique contribution to explaining coping, consequently avoiding collinearity among the overlapping emotions that might be included. Thus, the coefficients for the emotions in the final model were only for those emotions that were selected in the first stage because the others that were not selected made no significant further contribution. The emotions that made no additional contribution were omitted from the estimation equations in the final model. In this way, we avoided problems with collinearity among the many highly correlated emotions in the model estimation.

To test the robustness of this model, the equality of these coefficients across racial/ethnic groups were then tested, as was the equality of the coefficients across the four identities and gender. ² There were no significant differences across the four identities (the friend, romantic, worker, and family identities), χ²(76) = 74.98, p = .51; among the racial groups, χ²(318) = 315.89, p = .31; or between the genders, χ²(38) = 39.36, p = .41. Thus, the model was robust for the effects across important social categories of people. Because the effects of gender, race/ethnicity, or one’s identity had no effect on the outcomes, they were omitted in the final analyses. The final model showed a good fit of the model to the data (root mean square error of approximation = .00).