Sensemaking and Negative Emotion Sharing: Perceived Listener Reactions as Interpersonal Cues Driving Workplace Outcomes

Abstract

Emotions are a ubiquitous part of the workplace, and research on emotion sharing suggests that people often seek out others to express and share their emotions, in particular their negative emotions. Drawing from theory on sensemaking within organizations, we argue that employee perceptions of listener responses to negative emotion/stressful-event sharing have a significant impact on how included employees feel with their peers and their organization-based self-esteem. If employees perceive that others have responded positively to their sharing of negative emotions, they will experience positive inclusion and esteem beliefs and seek to maintain these positive views through socially attached attitudes (i.e., greater commitment and lower turnover intentions). However, if employees perceive listeners have responded negatively, inclusion, esteem, and socially attached attitudes will suffer. Across two studies, we found that employees’ perceptions that others tended to respond to them in a positive manner (e.g., by being supportive and validating the employees’ perspective) predicted the extent to which employees felt included, and experienced positive organization-based self-esteem. In turn, this translated into increased organizational commitment and decreased turnover intentions. In contrast, perceived negative listener reactions (e.g., responding in a critical or disengaged manner) threatened these outcomes. Our second study suggested that these effects are more driven by identity (i.e., esteem)-related processes than by generalized perceptions of social support, and that these effects are stronger for those with lower levels of communion striving.

Appendix

Pilot data on this measure were collected from 297 participants on Amazon’s Mechanical Turk (after data cleaning; total N before data cleaning = 409). We took recommended steps to ensure data quality, for instance by requesting participants with high approval ratings and by including checks for careless responding (Cheung, Burns, Sinclair, & Sliter, 2017). Half of the participants were female (51.4%), and on average, they were around 34 years old (SD = 10.17, range = 18–65 years). Participants reported an average yearly income of $40,838 (range = $3000–$225,000) and working 42.51 h a week (SD = 7.10, range = 15–85). The average time spent working for a current employer was about 5 years (SD = 4.92, range = < 1–32 years), and most participants were not serving in a management position (70.1%). Finally, most participants had a college degree (48.3%); 27.7% had some college education, 10.6% had either some high school or a high school degree, and 13% reported earning a master’s degree or some type of a professional degree.

The purpose of the pilot study was to identify the best-performing items to select and include as part of a brief measure of positive and negative listener reactions as well as to assess construct validity. We identified the highest loading items in an exploratory factor analysis using the .70 factor loading cutoff as the decision criteria (Tabachnick & Fidell, 2006). This resulted in 12 positive listener reaction items and 6 negative listener reaction items. Notably, the 12 highest loading positive listener reactions loaded across three different factors in the EFA while all 6 of the negative listener reactions loaded on a single factor. Recognizing the possibility that the positive listener reaction measure could be separated into three subscales (capturing emotionally oriented positive responses, informationally oriented positive responses, and validation), we conducted a confirmatory factor analysis using the 12 positive items and the 6 negative items. Findings indicated acceptable fit for the four-factor model (X²(129) = 300.16, p < .001; CFI = .94, RMSEA = .073 (.063, .084), SRMR = .047).

However, previous theory and research on listener reactions and interpersonal cues suggests that listener reactions of the same valence (e.g., listener reactions that are all generally positive) could be combined into a higher-order factor as they essentially reflect the same overall affirmation appraisal (e.g., Suhr et al., 2004; Wrzesniewski et al., 2003). Additionally, recent meta-analytic findings in the social support literature suggest that emotional and instrumental supports (related to the emotionally oriented and information-oriented positive listener reactions, respectively) are strongly related across the literature (ρ = .73; Mathieu et al., 2018), suggesting that these constructs are quantitatively similar. Since our own factor correlations indicated significantly strong correlations among the positive listener reaction factors (ranging from r = .45–.61), suggesting the possibility of an overall affirmation interpretation of the positive interpersonal cues, we tested a factor structure where the three lower-order positively valenced listener reactions were loaded onto a positive second-order factor. All negatively valenced items remained loaded onto the negative listener reaction factor (representing the overall disaffirmation cue). Findings indicated acceptable fit for the second-order model (X²(131) = 315.07, p < .001; CFI = .94, RMSEA = .075 (.065, .086), SRMR = .068). Thus, we maintained the second-order factor structure to best reflect both theoretical and methodological considerations of our measure of affirming interpersonal cues (i.e., positive listener reactions). The reliabilities for both the positive (α = .93) and negative (α = .90) listener reaction measures were satisfactory.