Pupillometry and Hindsight Bias: Physiological Arousal Predicts Compensatory Behavior. Willem W. A. Sleegers, Travis Proulx, Ilja van Beest. Social Psychological and Personality Science, November 10, 2020. https://doi.org/10.1177/1948550620966153
Rolf Degen's take: https://twitter.com/DegenRolf/status/1326117586048790529
Abstract: According to violation–compensation models of cognitive conflict, experiences that violate expected associations evoke a common, biologically based syndrome of aversive arousal, which in turn motivates compensation efforts to relieve this arousal. However, while substantial research shows that people indeed respond with increased arousal to expectancy violating events, evidence for the motivating role of arousal is rarely found. In two within-subjects studies (N = 44 and N = 50), we demonstrate evidence for the motivating role of arousal in this violation–compensation process among university students. Using pupillometry and the hindsight bias phenomenon, we show that people respond with greater arousal when presented with expectancy violating information. In turn, we show that the pupillary response is positively related to the amount of hindsight bias being displayed. These findings provide further insights into the process underlying the hindsight bias and, crucially, support key predictions following from threat–compensation models.
Keywords: threat–compensation, arousal–behavior link, pupillometry, hindsight bias
We aimed to demonstrate the first direct link between physiological arousal and compensatory behavior. While the results of each study separately were not conclusive, the results from both studies combined did provide evidence for this link. Greater pupil dilation in response to an unexpected correct answer was associated with more hindsight bias. That is, participants shifted their second answer more toward the factual question’s correct answer, relative to their first answer, when they showed a larger physiological response to the correct answer to the question. This compensatory response following increased arousal is consistent with violation–compensation theories (Jonas et al., 2014; McGregor et al., 2012), specifically with the shared assumption that inconsistencies evoke arousal that causes compensation reactions.
That expectancy violations induce a syndrome of aversive arousal is an important tenet of violation–compensation theories. There is abundant evidence for this first link between expectancy violations and arousal, whether the expectancy violation involves perceptual anomalies (Sleegers et al., 2015), cognitive dissonance (Gerard, 1967), self-view inconsistencies (Ayduk et al., 2012), worldview violations (Townsend et al., 2010), or category-based violations (Mendes et al., 2007). Evidence for the second link, between arousal and the subsequent compensatory behavior, is rarely observed and limited to indirect assessments of arousal such as self-report measures (Laurin et al., 2008; McGregor et al., 2013, Experiment 4; Plaks et al., 2005) and the misattribution of arousal paradigm (Kay et al., 2010; Losch & Cacioppo, 1990; Proulx & Heine, 2008; Zanna & Cooper, 1974). Our findings provide more direct evidence for the often postulated relationship between arousal and compensatory behaviors following expectancy violations.
Two reasons might explain why we were able to demonstrate a link between arousal and compensatory behavior. First, recent developments in eye tracker technology have made this technology exceptionally noninvasive. Consequently, an eye tracker is less likely to evoke arousal that interferes with the arousal process underlying violation–compensation reactions. Second, we repeatedly presented participants with an expectancy violation and an opportunity to compensate—a requirement for physiological measures to improve reliability.
Limitations and Future Research
In our studies, we relied on pupillometry to assess an aversive state of arousal following negative belief feedback because threat–compensation theories strictly postulate a state of aversive arousal to motivate subsequent compensatory behaviors. However, while pupillometry is a valid measure of physiological arousal, it is not a direct measure of aversive arousal (e.g., Bradley et al., 2008). We believe our findings nevertheless plausibly indicate a state of aversive arousal. Studies have shown that negative belief feedback and states of surprise are (at least initially) experienced as aversive (Hajcak & Foti, 2008; Noordewier & Breugelmans, 2013; Noordewier et al., 2016). In addition, alternative explanations such as curiosity-driven responses were ruled out by the data (see Online Appendix C). We therefore believe our findings present a strong contribution to models of threat–compensation.
It should be noted that we relied mostly on epistemic threats rather than more severe existential threats such as those relating to one’s identity or freedom. Epistemic threats were chosen in order to be able to repeatedly present participants with threats and compensation opportunities. This would not be feasible when more impactful threats are used because the physiological response would likely carry over between trials and affect the relationship between arousal and compensation. Moreover, the theoretical perspectives that guide this research share the explicit premise that the response to epistemic threats generalize to other types of threats (Heine et al., 2006; Jonas et al., 2014; Proulx & Inzlicht, 2012). In fact, it has been demonstrated that the experience of inconsistency, such as those experienced by our participants, can evoke the same compensation behaviors as existential threats (e.g., nonsense word pairs and identity violations; Randles et al, 2011). Nevertheless, the threat–compensation literature would benefit from more empirical demonstrations of the kind presented here.
Aside from expectancy violations inducing physiological arousal, and physiological arousal motivating compensatory behavior, compensatory behavior should also reduce the physiological arousal. We did not assess this third link. Using the present studies’ design, it might be possible to demonstrate the entire causal link by having participants again see the correct answers. We predict that instead of the positive relationship between pupil size and hindsight bias found in the present study, a negative relationship between hindsight bias and pupil size should be found.
Finally, in the present studies, we used the hindsight bias as a way to repeatedly assess compensatory behaviors following belief violations. It may be argued that due to the many trials, participants may not have always remembered their initial answer and that this ultimately shaped their hindsight bias responses. However, research on the hindsight bias largely supports a biased reconstruction view rather than a memory impairment process (Stahlberg & Maass, 1997). Our findings also contribute to the research on the hindsight bias. Several processes have been proposed to explain the hindsight bias (Hawkins & Hastie, 1990), including motivational accounts (Campbell & Tesser, 1983; Fischhoff, 1975; Musch, 2003). Our results are consistent with a motivational interpretation of the hindsight bias, thereby also contributing to research on the hindsight bias phenomenon.
We did employ a memory design to measure hindsight bias. Importantly, this memory-based design, although effective in demonstrating a hindsight bias, might be less effective in evoking a hindsight bias than other designs such as the hypothetical design (Pohl, 2007), in which participants are asked to respond as if they had not been told the correct answer. After all, a memory task is about recalling a previously reported answer; and when the time lag is not substantial, people can with relative ease recall their answer. For this reason, the memory design can be potentially improved in future studies by extending the retention interval between the first and second responses.