Testing an Ironic Process Model of Emotion Expression in Daily Life

Expressive suppression, or suppressing the expression of an emotion, is a common emotion regulation approach that has been found to have substantial negative implications for one’s mental health, physical health, and relationship functioning. However, expression of emotion, and of negative emotion in particular, has not consistently been found to be beneficial for mood or relationships. Moreover, the social expression of emotion that is culturally valued among Black-identifying Americans is different in White-identifying Americans, particularly in regard to expression of negative affect. However, it is unknown whether this cultural context mitigates the potential consequences of expressive suppression and emotion expression. Expressive suppression and expression of emotion have been established as distinct constructs rather than two ends of one continuum. However, whether expressive suppression serves as a context that influences the expression of negative emotion and the associated outcomes has not been thoroughly explored. The present study tested an ironic process effect of emotion expression, by which suppression of emotion may result in paradoxical expression of that emotion after a delay. Using 14 days of daily-diary surveys, the present study tested an ironic process model of emotion expression in emerging adulthood in daily life (Aim 1), examined whether expressive suppression moderated associations between emotion expression and outcomes for mood and relationships (Aim 2) and whether these effects varied across White and Black Americans (Aim 3). Results indicated that expressive suppression was not directly associated with intensity of negative emotion expression, failing to support the hypothesis in Aim 1. However, results did reveal a significant interaction between expressive suppression and intensity of negative emotion expression in models predicting emotional intimacy, both within and between participants. Between participants, the effect was in the direction hypothesized in Aim 2, but within-participants, the direction of the effect was reversed. Similarly, in the between-person model predicting anxiety symptoms, the interaction was significant, but not in the direction that would be consistent with an ironic process effect. In the model predicting depression symptoms, the interaction was not significant within or between participants after controlling for race. Cumulatively, these results provide partial support for the hypothesized ironic process effect in Aim 2. Lastly, there was no support for a three-way interaction between racial identity and the ironic process effect, suggesting that effects did not vary across race. However, in models predicting depression and anxiety, the positive association between expression of negative emotion and mood was moderated by race such that there was no significant association in participants who identified as Black. Clinical implications and future directions are discussed.