The Associations Between Emotions, Evaluations, and Risk

Revisiting seminal research on the classic appraisal-tendency framework.

Reviewed by Lybi Ma

Key points

• The appraisal-tendency framework suggested that emotions are associated with judgments in different ways.
• Framework views emotions as more complex than mere positive-negative valence view.
• We revisited a seminal article by Lerner and Keltner (2000) with a replication and extension of their work.
• We found support for some key predictions, yet not for all, suggesting a more nuanced view.

By Sirui Lu

Each emotion has its functions, yet way too often we classify emotions based on their valences, positive emotions are good, negative emotions are bad. This valence-based classification not only applies to laypersons but also was the early approach to studying affect’s role on people’s cognitions (DeSteno and colleagues, 2000; Lerner and Keltner, 2000). An alternative, more nuanced view proposed that specific emotions are associated with judgment and decision-making in different ways, even if they might be of the same valence.

Appraisal-tendency framework

Lerner and Keltner (2000) proposed an appraisal-tendency framework. They argued that different emotions are associated with different cognitive appraisals, such as assessments regarding level of control and certainty, that in turn may predict one’s evaluations of risk and events. For example, they compared anger and fear, both considered negative in valence, and found that people who have a higher propensity for anger (high on trait anger) tended to make more optimistic future judgments and preferred risk-seeking decisions, whereas those who have a higher propensity for fear (high on trait fear) tended to make more pessimistic judgments and preferred risk-aversive decisions (Lerner and Keltner, 2000, 2001). They explained the deviations as rooted in the controllability and certainty dimensions of emotions, two of the six cognitive dimensions of emotions proposed in the appraisal-tendency framework (the other four are pleasantness, anticipated effort, attentional activity, and responsibility). Consistent with the framework, their studies on anger (negative, high certainty/control), fear (negative, low certainty/control), and happiness (positive, high certainty/control) demonstrated that trait anger and trait happiness similarly indicated optimism, whereas trait fear was associated with pessimism. Furthermore, ambiguity moderated these effects, the valence-based approach matched better with unambiguous events (Lerner and Keltner, 2001).

This seminal article has had an impact on scholarly research in the areas of judgment, decision-making, and emotion. Given that over two decades have passed, with no direct independent replication, we thought it was time to revisit this phenomenon and conduct a replication with some new extensions.

Our design and findings

We conducted a close replication with a large sample and made some needed methodological improvements with added extensions to try and gain new insights. In addition to the three specific emotion traits examined in the original study, we added trait hope that completes the emotions range with a positive valence of low certainty/control (see Table 1). According to the appraisal-tendency framework, we expected that trait hope would show a pattern similar to that of happiness in predicting unambiguous events, yet showing a similar pattern to fear in predicting ambiguous events.

Our study had a four (dispositional emotions: anger, fear, happiness, hope) by two (ambiguity of events: ambiguous events and unambiguous events) within-subject design, examining risk optimism and risk preferences. We ran the study with 780 US American participants.

We successfully replicated some of the original findings, finding support for several key predictions. We found that trait anger and trait happiness were positively associated with risk-seeking and optimistic risk estimates. Trait fear was negatively associated with optimistic risk assessment though unexpectedly not with risk-seeking.

Our extension of adding trait hope did not work out as expected, inconsistent with our predictions that were based on the appraisal-tendency framework. One of the many possible reasons could be that research is still nascent when it comes to some specific emotions, including trait hope. For example, some researchers suggested that hope carries mixed and complex feelings of distress and worry (Halevy, 2017) and is distinct from other positive emotions (Rosler and colleagues, 2016).

We hope to see future research test more emotions, not only in terms of valences and controllability/certainty but also consider including other aspects of emotions and their influence on a broader range of decision-making processes.