(PS1-100) Longitudinal effect of anger suppression and expression on social support: Trait anger and anger-in predict decreased social support over time.

Background: Conventional wisdom holds that excessive anger as well as emotion suppression are harmful to one’s health (Mauss & Gross, 2004). Evidence suggests that anger suppression may intensify anger and impede problem solving (Larsson et al., 2023). Emotion research has long held that expressing emotions is beneficial, activating the attachment system and promoting interpersonal bonding (Rime, 2009). However, evidence has emerged that expressing negative feelings may worsen social outcomes, especially in the case of anger expression, which may be interpreted as a threat (Chervonsky & Hunt, 2017). Another hypothesis holds that anger expression may be a form of emotion suppression, in that it can suppress more primary emotions of shame or sadness (Larsson et al., 2023). The present study investigated the longitudinal effects of anger suppression and expression as measured by the STAXI on social support at a ~9-year follow up in the Midlife in the US (MIDUS) dataset.

Methods: Using longitudinal data from the MIDUS 2 and 3 projects, this study included 884 adults who completed self-report measures of anger expression and social support. The State-Trait Anger Expression Inventory (STAXI) (Spielberger et al., 1995) assessed trait anger (TA) as well as anger expression, with subscales measuring tendencies toward Anger-In (AI) and Anger-Out (AO). Four social support variables included: i) a Berkman-Syme-type social network index constructed from reported frequency of different types of social interactions (Glei et al., 2012); ii) a perceived social support scale constructed from three MIDUS scales of perceived support from partner, family, and friends (Elliot et al., 2018); iii) a MIDUS scale of self-reported perceived social integration; and iv) a MIDUS scale of perceived positive relations with others. We regressed time 2 (T2, MIDUS 3) social support variables on time 1 (T1, MIDUS 2) anger variables (controlling for T1 social support) to assess whether anger variables longitudinally predicted residualized change in social support. The interval between T2 and T1 data collection ranged from 7.9 to 10.3 years, with a mean of 9.1 years.

Results & Conclusions: Multiple regression analyses revealed significant negative associations of T1 TA and AI with T2 social support variables. Higher T1 TA predicted lower T2 social support (β = -0.12, p < .001) and positive relations (β = -0.12, p < .001). Higher T1 AI predicted lower T2 social integration (β = -0.07, p = .02), social support (β = -0.12, p < .001), and positive relations (β = -0.11, p < .001). T1 AO did not significantly predict T2 social support outcomes. TA did not moderate the relations between T1 anger suppression/expression and T2 social support. Interestingly, these results suggest that “bottling up” one’s anger might tend to push others away while openly expressing anger does not. It may be that the adaptive social functions of outward anger expression may counteract its social aversiveness. These findings provide support to the well-established hypothesis that anger suppression is a maladaptive strategy.