The role of self-aspects in emotions elicited by threats to physical health

Abstract

In two studies, we examined the relationship between self-aspects and socially engaging and socially disengaging emotions elicited by imagined and real physical health problems. In Study 1, participants imagined themselves experiencing a health problem described in a hypothetical scenario and rated the extent to which they would experience a list of emotions. The experience of socially engaging emotions such as shame and embarrassment was predicted by the endorsement of collective self. In Study 2, participants recalled a past health problem and emotions they experienced during its course. Again, collective self predicted the extent to which people mentioned socially engaging emotions in their free recall of emotions. Independent self was not related to the imagined experience of socially disengaging emotions in Study 1 or the recollection of such emotions in Study 2.

Keywords:

• self-aspects
• collective self
• illness-related emotions

Acknowledgements

We would like to thank Lynda Cheng, Adele Ambrose, Uzma Jeelani, Andy Ng, Mandeep Singh, Marina Veprinska and Syeda Abedi for their assistance in conducting the studies and coding the open-ended data. We also thank Keiko Ishii for her helpful comments on the manuscript. This research was supported by a doctoral fellowship from the Ontario Women’s Health Council, Canada to Ayse Uskul. Uskul was funded by a postdoctoral fellowship of the Social Science and Humanities Research Council of Canada while working on the manuscript.

Notes

1. The positive emotions factor (α = .72, M = 3.68, SD = 1.11) consisted of five emotions (see Table 1) and negatively correlated with socially disengaging emotion factor (r = −.20, p = .001) only.

2. When all possible two-way interactions were entered in Step 2 and all possible three-way interactions were added in Step 3, no additional variance was explained.

3. Although we had no predictions for positive or neutral emotions, we also conducted a regression analysis with the emotion factor that included positive and neutral emotions. This regression revealed a significant R² at the end of Step 1 only,  = .04, F∆ (5263) = 3.11, p = .01. The only significant predictor was collective self-aspect (RIC-C), β = .17, t (263) = 2.39, p = .02.

4. Participants wrote on a diverse set of health problems ranging from bronchitis to cancer. The diversity of health problems did not allow for a meaningful categorisation in terms of type of health problem that could be treated as an additional factor in the analyses.

5. Participants completed both independent and interdependent subscales of Singelis’ (1994) self-construal Scale. The interdependent subscale in this study had an unacceptable reliability coefficient and was found to have a multifactorial structure. Given that the scale was not measuring a clear construct that could be labelled as the ‘interdependent self-construal’, we decided to report the independent self-construal subscale only. However, when we included the interdependent subscale into our analyses, we found that the findings remained unchanged and that this variable was not a significant predictor.

6. Even though we recruited participants with health problems that lasted at least 2–3 days, two participants who wrote down a 1-day long health problem. They were not excluded from the sample since their inclusion or exclusion did not alter the results.

7. The positive emotions factor (α = .72, M = 3.44, SD = 1.27) consisting of the emotions confident, hopeful, trusting, strong and responsible (see Table 1) was only marginally negatively correlated with socially disengaging emotion factor (r = −.20, p = .08).

8. We referred to the reasons for the inclusion of the closed-ended emotions scale in this study in the description of this scale above. Having these scales included in this study gave gives us an opportunity to examine whether the observed pattern of findings in Study 1 can be replicated in a different sample and conducted similar regression analyses to those in Study 1. Unfortunately, however, we did not find that the Singelis independent self-construal subscale predicted socially disengaging emotions (as in Study 1 with the RIC-I) and that the collective self construal scale predicted socially engaging emotions (unlike the RIC-C in Study 1). The reasons for the lack of prediction may lie in the self-aspects scales used in this study. The reliability of the independent self-construal subscale was less than desired and the collective self-construal scale tapped primarily into the individuals’ group related selves, as opposed to interrelated selves, which may have been less relevant in the context of the recalled health problems.

9. The regression analyses reported in this section were repeated controlling for duration, seriousness of the health problem, age and having children. Since results remained the same with the inclusion of these demographic variables, we report the results of regressions that include the variables of interest only.