Can people detect the trustworthiness of strangers based on their facial appearance? It seems not.

Can people detect the trustworthiness of strangers based on their facial appearance? Bastian Jaeger et al. Evolution and Human Behavior, May 2 2022. https://doi.org/10.1016/j.evolhumbehav.2022.04.004

Abstract: Although cooperation can lead to mutually beneficial outcomes, cooperative actions only pay off for the individual if others can be trusted to cooperate as well. Identifying trustworthy interaction partners is therefore a central challenge in human social life. How do people navigate this challenge? Prior work suggests that people rely on facial appearance to judge the trustworthiness of strangers. However, the question of whether these judgments are actually accurate remains debated. The present research examines accuracy in trustworthiness detection from faces and three moderators proposed by previous research. We investigate whether people show above-chance accuracy (a) when they make trust decisions and when they provide explicit trustworthiness ratings, (b) when judging male and female counterparts, and (c) when rating cropped images (with non-facial features removed) and uncropped images. Two studies showed that incentivized trust decisions (Study 1, n = 131 university students) and incentivized trustworthiness predictions (Study 2, n = 266 university students) were unrelated to the actual trustworthiness of counterparts. Accuracy was not moderated by stimulus type (cropped vs. uncropped faces) or counterparts' gender. Overall, these findings suggest that people are unable to detect the trustworthiness of strangers based on their facial appearance, when this is the only information available to them.

Keywords: TrustTrustworthinessCooperationFace perceptionPredictionAccuracy

4. General discussion

People spontaneously rely on the facial appearance of strangers when deciding whether they can be trusted to cooperate in social interactions (Todorov, Olivola, et al., 2015). But can people actually detect the trustworthiness of strangers based on their facial appearance? Prior studies have yielded mixed results and the question remains the subject of vigorous debate (Bonnefon et al., 2017; Todorov, Funk, & Olivola, 2015). Yet, the empirical evidence on the topic is limited. Many studies were based on the same set of stimuli, which limits the generalizability of findings (Bonnefon et al., 2013; De Neys, Hopfensitz and Bonnefon, 2015, De Neys, Hopfensitz and Bonnefon, 2017). Conversely, studies providing evidence against accuracy relied on statistical techniques that cannot quantify evidence in favor of such a null hypothesis, which complicates the interpretation of results (Efferson & Vogt, 2013; Rule et al., 2013).

We conducted two studies to address these limitations. Confirming results from previous studies (Jaeger, Evans, Stel, & van Beest, 2019), we found that participants relied on the perceived trustworthiness of counterparts when making trust decisions. However, on average, participants failed to entrust money to counterparts that were actually more trustworthy. Bayesian analyses yielded very strong support for the null hypothesis indicating that our participants were not able to accurately detect the trustworthiness of their interaction partners. We also found that participants' earnings were not higher than the expected earnings of a decision strategy that trusts at random. This suggests that knowledge of their counterparts' facial appearance did not give participants a strategic advantage. In fact, participants would have earned more by consistently distrusting all counterparts, as trust did not pay off in the current sample.

Previous studies found evidence in favor of detection accuracy only under specific conditions, and these conditions varied across studies (Bonnefon et al., 2013; Tognetti et al., 2013; Verplaetse et al., 2007). Here, we tested these proposed moderators, but found no evidence for better-than-chance trustworthiness detection (a) for male or female counterparts, (b) when making trust decisions or when providing explicit trustworthiness ratings, and (c) when viewing cropped images (in which all non-facial features were removed) or uncropped images. In sum, our results provide consistent evidence against accuracy in trustworthiness detection from faces across various conditions.

Previous investigations have shown that trustworthiness impressions guide decision-making in many domains, including legal sentencing, personnel selection, and financial decision-making (Olivola et al., 2014). People even rely on trustworthiness impressions from faces when more diagnostic cues are available (Jaeger et al., 2019) and when decisions are highly consequential (Wilson & Rule, 2015). Future studies should explore whether some people are more prone to the biasing influence of first impressions and, importantly how biases could be mitigated (for first attempts, see Chua & Freeman, 2021; Jaeger, Todorov, Evans, & van Beest, 2020; Shen & Ferguson, 2021). An important future task in this line of research will be to delineate how difficult it is to override these biases, particularly when other more reliable information sources are available that may require more cognitive effort to process.

4.1. Limitations and future directions

Several limitations and constraints on the generalizability of the current results should be mentioned. Our results were based on samples of relatively young decision-makers from the University of Zurich. Additional studies are needed to examine the generalizability of our findings with larger and more diverse samples of both targets and raters.

Future studies should also examine the accuracy of trustworthiness impressions using varying types of stimuli. Cropped images, in which all non-facial aspects are removed, ensure that impressions are based on the facial features of counterparts. However, they represent only a relatively specific facet of the kinds of stimuli that people encounter in everyday interpersonal interactions. Accuracy may be better than chance when people have access to additional cues. For instance, previous findings suggest that people may be able to identify cooperative interaction partners with greater-than-chance accuracy after brief interactions (Brosig, 2002; DeSteno et al., 2012; Frank, Gilovich, & Regan, 1993; Reed, Zeglen, & Schmidt, 2012; but see Manson, Gervais, & Kline, 2013; McCullough & Reed, 2016). Ultimately, we believe that studies using a wide range of different stimuli are needed to map the accuracy of trustworthiness decisions under varying conditions.

We investigate one such condition here, namely the accuracy of trustworthiness judgments when judgments are solely based on facial features. This approach is informative for two reasons. First, even though people often have access to other cues, which may allow them to make more accurate judgments, there are also many situations in which a person's facial appearance is either one of the only cues or a particularly salient cue. People often engage with strangers and, in the first moments of the interaction, tend to judge them solely based on their appearance. Moreover, facial photographs are a common feature of many decision-making environments, including social media platforms (e.g., Twitter), professional networking sites (e.g., LinkedIn), and the sharing economy (e.g., Airbnb). Second, ample evidence suggests that people rely on facial appearance, even when they have access to other cues (Jaeger et al., 2019; Olivola et al., 2014). To determine whether reliance on facial appearance helps or hinders people in making accurate predictions, the accuracy of judgments that are solely based on facial appearance needs to be isolated. This requires a highly controlled and standardized study design, such as the one used in the current experiments, to ensure that judgments are based on the cue in question (Cox et al., 2015).