Immediate emotions and subjective stakes in risky decision-making under uncertainty

ABSTRACT

Background

Previous research has shown that immediate emotions and cognitive processing of the stakes of outcomes influence decision-making under uncertainty. The effect of perceived beneficial stakes and different types of immediate emotions on decision-making is an important topic that has received little attention in the literature. This study investigated the effects of trait anxiety and anticipatory emotions (fear, sadness, excitement and comfortability) on the perception of thee stakes of outcomes and behavioral intentions.

Method

Participants from the community completed a task measuring anticipatory emotions and their perceived stakes of risky and beneficial outcomes in a range of uncertain situations. Trait anxiety was also measured.

Results

Results revealed that anticipatory emotions (except for sadness), trait anxiety and subjective stakes all demonstrated significant associations with risky behavioral intention in uncertain situations. Anticipatory emotions, but not trait anxiety, had stable effects on stake perceptions. However, trait anxiety moderated the effect of excitement on risky behavioral intention. In addition, positive emotions (comfortability and excitement) and beneficial stakes demonstrated consistent effects in the decision-making process.

Conclusions

The current study sheds light on future immediate-emotion-based interventions for deficits in uncertain decision-making.

KEYWORDS:

• Anticipatory emotions
• immediate emotions
• risk
• decision-making
• anxiety
• uncertainty

Disclosure statement

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

Data sharing

Data associated with this study are available upon request.

Notes

1 To our knowledge, there was no well-established methods to compute power for the mixed-effects ordinal regression we were applying. G*Power using fixed-effect logistic regression as an approximation indicates that a minimum of 881 observations are required for detecting a small effect size (Odds ratio = 1.22, Olivier, May, & Bell, 2017) at a power of 0.9 and a significance level of 0.05. Ali, Ali, Khan, and Hussain' (2016) simulation study indicates that 50 groups (participants) with a group size of 30 (30 responses per participant) would lead to a power over 0.85 for five-category mixed-effects ordinal model. In addition, from the viewpoint of having accurate parameter estimation, a minimum of 420 observations (20:1 data-point to parameter ratio) are required for estimating fixed-effect parameters in the largest model in H4 (9 thresholds and 12 predictors). The current sample of 80 participants that provided 2400 observations in total (30 responses per person) would be sufficient.

Additional information

Funding

This research was supported by the Australian Government through the Australian Research Council (Project number DE180100015).