Examining effects of anticipated stigma, centrality, salience, internalization, and outness on psychological distress for people with concealable stigmatized identities

Abstract

Understanding how stigmatized identities contribute to increased rates of depression and anxiety is critical to stigma reduction and mental health treatment. There has been little research testing multiple aspects of stigmatized identities simultaneously. In the current study, we collected data from a diverse, urban, adult community sample of people with a concealed stigmatized identity (CSI). We targeted 5 specific CSIs--mental illness, substance abuse, experience of domestic violence, experience of sexual assault, and experience of childhood abuse--that have been shown to put people at risk for increased psychological distress. We collected measures of the anticipation of being devalued by others if the identity became known (anticipated stigma), the level of defining oneself by the stigmatized identity (centrality), the frequency of thinking about the identity (salience), the extent of agreement with negative stereotypes about the identity (internalized stigma), and extent to which other people currently know about the identity (outness). Results showed that greater anticipated stigma, greater identity salience, and lower levels of outness each uniquely and significantly predicted variance in increased psychological distress (a composite of depression and anxiety). In examining communalities and differences across the five identities, we found that mean levels of the stigma variables differed across the identities, with people with substance abuse and mental illness reporting greater anticipated and internalized stigma. However, the prediction pattern of the variables for psychological distress was similar across the substance abuse, mental illness, domestic violence, and childhood abuse identities (but not sexual assault). Understanding which components of stigmatized identities predict distress can lead to more effective treatment for people experiencing psychological distress.

Figure 1. Interaction between Centrality and Internalization on Psychological Distress.

At most levels of identity centrality (below the 75^th percentile), the relationship between internalization and psychological distress is not significant. At very high levels of centrality, however, greater internalization is related to greater psychological distress. Note: Plot based on predicted values of distress given values of the predictor variables at the 10^th, 25^th, 50^th, 75^th, and 90^th percentiles. Predictor variables centered.

Figure 2. Means and Standard Errors for Anticipated Stigma for each CSI group.

Analyses controlling for income, education, and sex found significant differences in mean levels of anticipated stigma reported across the groups. Bonferroni corrected post-hoc comparisons show that the substance abuse groups reported greater levels of anticipated stigma than the child abuse (p<.001), sexual assault (p<.001), and domestic violence groups (p = .03) but not more than the mental illness group. Similarly, the mental illness group reported more anticipated stigma than the child abuse (p<.001) and sexual assault groups (p = .003) but not the domestic violence group.

Figure 3. Means and Standard Errors for Internalization for each CSI group.

Analyses controlling for income, education, and sex found significant differences in mean levels of internalization of stigma reported across the groups. Bonferroni corrected post-hoc comparisons show that the substance abuse groups reported greater levels of internalization than the child abuse (p = .001) and sexual assault groups (p = .003).

Figure 4. Means and standard errors for Centrality for each CSI Group.

Analyses controlling for income, education, and sex found significant differences in mean levels of centrality reported across the groups. Bonferroni corrected post-hoc comparisons show that only the substance abuse and mental illness group report significantly different levels of centrality (p = .005). No other groups differ.

Figure 5. Means and standard errors for Salience for each CSI Group.

Analyses controlling for income, education, and sex found significant differences in mean levels of salience reported across the groups. Bonferroni corrected post-hoc comparisons show that the mental illness group reports significantly higher identity salience than the childhood abuse (p<.001), sexual assault (p = .002), and domestic violence groups (p = .001), but no difference from the substance abuse group. Likewise, the substance abuse group differs from the childhood abuse (p = .003) and domestic violence groups (p = .03) although it is only marginally different from the sexual assault group (p = .07).

Figure 6. Means and standard errors for Outness for each CSI Group.

Analyses controlling for income, education, and sex found significant differences in mean levels of outness reported across the groups. Bonferroni corrected post-hoc comparisons show that the substance abuse group reported being significantly more out about their identity than the childhood abuse (p<.001), sexual assault (p = .001), and domestic violence (p = .001) groups but not different from the mental illness group (p = .12). There were no other significant between group differences.