Fast Foes: the physiological and behavioral consequences of interacting in an immersive negative social context

ABSTRACT

Background and Objectives

Social interactions involving personal disclosures are ubiquitous in social life and have important relational implications. A large body of research has documented positive outcomes from fruitful social interactions with amicable individuals, but less is known about how self-disclosing interactions with inimical interaction partners impacts individuals.

Design and Methods

Participants engaged in an immersive social interaction task with a confederate (thought to be another participant) trained to behave amicably (Fast Friends) or inimically (Fast Foes). Cardiovascular responses were measured during the interaction and behavioral displays coded. Participants also reported on their subjective experiences of the interaction.

Results

Participants assigned to interact in the Fast Foes condition reported more negative affect and threat appraisals, displayed more negative behaviors (i.e., agitation and anxiety), and exhibited physiological threat responses (and lower cardiac output in particular) compared to participants assigned to the Fast Friends condition.

Conclusions

The novel paradigm demonstrates differential stress and affective outcomes between positive and negative self-disclosure situations across multiple channels, providing a more nuanced understanding of the processes associated with disclosing information about the self in social contexts.

KEYWORDS:

• Social interactions
• self-disclosure
• Fast Friends
• biopsychosocial
• challenge
• threat

Acknowledgement

We would like to thank Hannah Gravelding for her help scheduling participants, and training experimenters and confederates, and the many research assistants from the Social Stress Lab who helped collect, process, and analyze data.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 The Fast Friends paradigm is ∼45-min and uses 36 questions. We reduced these to maximize control. The task was reduced to 5-min and used 24 of the original 36 questions.

2 We compared multilevel models that included the random intercept variance with models not including the random intercept variance using a LR ratio test for each of the DVs. Models differed for one of the twelve outcomes (Resources, LR chi2(1) = 7.08, p = .004). We ran an additional multilevel model allowing slopes to vary based on confederate assignment, and observed that the Condition main effect still manifested, t(179) = −2.62, p = .01.

3 A ladder-of-powers approach was used to determine an appropriate transformation. We first used a log tranformation, but the Shapiro-Wilks test revealed the data remained non-normal. Next, we used the reciprocal root, which reduced skew to a more acceptable value (p > .05). However, because the log and recipricol root transforms yielded similar results, we opted for the log tranformation.

4 To correct for multiple comparisons, the false discovery rate (FDR; Benjamini et al., 2001; Cramer et al., 2016) was controlled for. All but one effect (Intrapersonal NA) survived the correction.