Framing effects refer to how making one dimension of a decision problem salient influences subsequent choices. For example, one classic study revealed that negotiators who framed negotiations in terms of gains used more concessionary processes (and had more successful outcomes) than negotiators who framed negotiations in terms of losses (Neale & Bazerman, Reference Neale and Bazerman1985). Over the past half century, such framing effects are among the most widely documented and influential findings in the social and behavioral sciences (Ruggeri et al., Reference Ruggeri, Alí, Berge, Bertoldo, Bjørndal, Cortijos-Bernabeu and Folke2020; Tversky & Kahneman, Reference Tversky and Kahneman1981). Traditionally, although with important exceptions (e.g., Gigerenzer & Gaissmaier, Reference Gigerenzer and Gaissmaier2011; Sher & McKenzie, Reference Sher and McKenzie2006), framing effects are considered irrational biases because they violate the statistical axiom of invariance. In his target article, Bermúdez extends prior research by arguing that the role of frames in reasoning can give rise to rational framing effects. Here, I argue for the existence of rational framing effects for a complementary reason: Frames can shift reputational incentives for decision makers.

Both traditional research and Bermúdez's extension focus primarily on individual cognition. Yet, to assess when framing effects are rational, researchers need to consider the full set of goals individuals hold. Most relevant to the present article, individuals hold strong impression management goals. A large and interdisciplinary body of research studies reputation as a cause of behavior, arguing that attention to reputation shapes judgment and decision making (Lerner & Tetlock, Reference Lerner and Tetlock1999; Tetlock, Reference Tetlock2002). For example, prior research has made clear that accountability (e.g., making decisions in public vs. private) systematically shifts behavior (Lerner & Tetlock, Reference Lerner and Tetlock1999). To fully assess the rationality of framing effects, researchers must therefore broaden the scope of analysis to include reputational incentives. Specifically, researchers must examine the reputational consequences of framing effects.

To illustrate, consider perhaps the most famous example of framing effects: loss–gain framing effects on risk preferences (Kahneman & Tversky, Reference Kahneman and Tversky1979; Tversky & Kahneman, Reference Tversky and Kahneman1981). The canonical finding is well-documented: Decision makers are more likely to make risk-seeking choices when the potential outcomes of alternative choices are framed as losses. In contrast, they are more likely to make risk-averse choices when outcomes are framed as gains, even when choice sets are otherwise equivalent. Could it be the case that loss versus gain frames also shape which choice is perceived more positively by observers?

Recent advances support this theorizing. In a series of three online experiments, Dorison and Heller (Reference Dorison and Heller2022) found that third-party observers penalize decision makers whose risk preferences are unaffected by loss–gain framing. Specifically, observers reputationally punished leaders who made risk-averse (vs. risk-seeking) decisions when outcomes were framed as losses. In contrast, this result reversed when outcomes were framed as gains: risk-seeking leaders were punished. This reversal was robust across a variety of social dimensions (e.g., warmth vs. competence), choice contexts (e.g., public health decisions vs. monetary gambles), and even with financial stakes on the line. These effects occurred because observers themselves fell victim to loss–gain framing effects, and in turn reputationally derogated decision makers who did not. Of note, such patterns are not limited to framing effects: Across contexts, violating traditional prescriptions for rationality can yield reputational benefits, casting doubt on whether such violations should always be considered irrational in the first place (e.g., Cao, Kleiman-Weiner, & Banaji, Reference Cao, Kleiman-Weiner and Banaji2019; Dorison, Umphres, & Lerner, Reference Dorison, Umphres and Lerner2022; Jordan, Hoffman, Nowak, & Rand, Reference Jordan, Hoffman, Nowak and Rand2016; Tenney, Meikle, Hunsaker, Moore, & Anderson, Reference Tenney, Meikle, Hunsaker, Moore and Anderson2019).

When should framing effects thus be considered rational? It may depend on the balance between reputational incentives (e.g., appearing trustworthy) and non-reputational incentives (e.g., maximizing profits for an organization). Rational framing effects are more likely to occur in contexts where differences in reputational incentives are large and differences in non-reputational incentives are small. Consider a public leader who is equivocal about the effects of a COVID-19 policy, but knows that presenting a risky, negatively framed choice to the public would bolster her approval ratings before an upcoming election. The further away the election, the less a framing effect can be considered rational. Conversely, rational framing effects are less likely to occur when differences in reputational incentives are small, but differences in non-reputational incentives are large. Consider a low-profile CEO of a start-up, whose unpopular risk-seeking decisions avoid the public gaze but are critical for the success of her budding company. The more public the decisions, the more a framing effect can be considered rational. Future research is needed to tease apart how such incentives can be balanced and integrated into formal decision models.

Finally, understanding the reputational consequences of framing effects sheds light on new paths for behavior change. Prior work has identified strategies focused on training individual leaders (Morewedge et al., Reference Morewedge, Yoon, Scopelliti, Symborski, Korris and Kassam2015). To the extent that decision makers are attuned to reputational incentives, a complement to such a cognitive approach could involve examining – and shifting – reputational incentives. Rational framing effects may be more versus less likely to occur depending on the broader social or organizational context in which they take place. For example, a leader could set an organizational culture that values identifying, in the moment, how a choice could be framed differently (for related work, see Daniels & Zlatev, Reference Daniels and Zlatev2019). Future research is needed on this possibility.

Bermúdez provided an important advance in identifying how the role of frames in reasoning can yield rational framing effects. Here, I suggest that broadening the scope of analysis to consider reputational incentives provides a complementary reason for rational framing effects. Looking ahead, critical and exciting questions remain regarding boundary conditions for when framing effects should be considered rational and how irrational framing effects can be reduced. Answering these questions is essential to understanding not only rational framing effects, but also rationality in judgment and decision making generally. Applied applications abound for leadership, organizations, and public policy.