Kenneth Savitsky’s research while affiliated with Williams College and other places

The more similar the members of a group are to one another, the less reliable their collective judgments are likely to be. One way for individuals to respond to negative feedback from a group may thus be to adjust their perceptions of the group's homogeneity, enabling them to dismiss the feedback as unreliable. We show that individuals appreciate this logic (Study 1) and that they put it to strategic use by regarding the members of a group as more homogenous when the group judges them negatively than when it judges them positively (Studies 2, 3, and 4). We underscore the self-protective nature of this tendency by showing that individuals adjust their perceptions of a group's homogeneity more when they themselves are the target of the group's judgment than when the group judges someone else (Study 4).

Across three studies, members of underrepresented groups felt that they were the center of others’ attention when topics related to their group were discussed, and this experience was accompanied by negative emotions. Black participants reported that they would feel most ‘‘in the spotlight’’ when they were the only Black individual in a class in which the professor drew attention to their group with a provocative comment (Study 1). Black and Latino/Latina (Study 2) and female (Study 3) participants likewise reported that two confederates looked at them more when they heard (and believed the confederates had also heard) a recording that pertained to their group than when they heard a recording on a neutral topic—despite the fact that the confederates’ gaze did not differ across conditions. We discuss these results in light of research on solo status and targeted social referencing.

Individuals sometimes try to convey different identities to different people simultaneously or to convey certain information to one individual while simultaneously concealing it from another. How successfully can people solve these multiple audience problems and how successfully do they think they can? The research presented here corroborates previous findings that people are rather adept at such tasks. In Study 1, participants who adopted different identities in preliminary interactions with two other participants (acting the part of a studious nerd with one and a fun-loving party animal with the other) were able to preserve these identities when they interacted subsequently with both individuals at the same time. In Study 2, participants were able to communicate a secret word to one audience while simultaneously concealing it from another. Despite their skill at these tasks, however, participants in both studies were overconfident in their abilities, believing that they were better able to solve these multiple audience problems than they actually were.

People commonly believe they have contributed more to collaborative tasks than others give them credit for. We distinguish between two types of contributions – additions (such as adding words to a co-authored paper) and deletions (such as removing extraneous words) – and show that individuals are especially prone to receive less credit from others than they believe they deserve when their contributions consist of taking something away rather than adding something. Participants who shortened some writing believed they improved it just as much as did participants who lengthened some, but were seen by others as having contributed less. Although one can hardly fail to notice one's own deletions, these contributions – like any contributions that, by their very nature, leave little trace of themselves – are easy for others to overlook.

People commonly believe that they communicate better with close friends than with strangers. We propose, however, that closeness can lead people to overestimate how well they communicate, a phenomenon we term the closeness-communication bias. In one experiment, participants who followed direction of a friend were more likely to make egocentric errors—look at and reach for an object only they could see—than were those who followed direction of a stranger. In two additional experiments, participants who attempted to convey particular meanings with ambiguous phrases overestimated their success more when communicating with a friend or spouse than with strangers. We argue that people engage in active monitoring of strangers’ divergent perspectives because they know they must, but that they “let down their guard” and rely more on their own perspective when they communicate with a friend.

Prior research has found that people tend to overestimate their relative contribution to joint tasks [e.g., Ross, M., & Sicoly, F. (1979). Egocentric biases in availability and attribution. Journal of Personality and Social Psychology, 37, 322-336]. The present research investigates one source of this bias, and in doing so, identifies an important moderator of the effect. Three studies demonstrate that when people estimate their relative contribution to collective endeavors they focus on their own contribution and give less consideration to the contribution of their collaborators. This can cause overestimation for tasks in which total contributions are plentiful, but underestimation for tasks in which total contributions are few--despite the fact that both tasks reflect positively on the person who performs them. These results extend Ross and Sicoly's (1979) original analysis of bias in responsibility judgments, but also suggest that the tendency to overestimate one's relative contribution to collaborations is not as ubiquitous as once thought.

People have more information about themselves than others do, and this fundamental asymmetry can help to explain why individuals have difficulty accurately intuiting how they appear to other people. Determining how one appears to observers requires one to utilize public information that is available to observers, but to disregard private information that they do not possess. We report a series of experiments, however, showing that people utilize privately known information about their own past performance (Experiments 1 and 2), the performance of other people (Experiment 3), and imaginary performance (Experiment 4) when intuiting how they are viewed by others. This tendency can help explain why people's beliefs about how they are judged by others often diverge from how they are actually judged.

People tend to underestimate their ability to feel good when bad things happen to them, and underestimate the speed with which they will make their recovery. These tendencies have been documented in recent research on affective forecasting and are recounted in an excellent and engaging new book.

People tend to overestimate their contribution to joint tasks, in part because their own contributions are more memorable than the contributions of their collaborators. We examined some of the interpersonal consequences of this bias. Participants engaged in either a hypothetical (Experiment 2) or real (Experiment 1) cooperative task and learned how their collaborator ostensibly allocated responsibility. We varied how much credit the collaborator took for herself, and also how much credit she gave to the participant, factors confounded in past research. In each experiment, collaborators who stole the glory were seen as less fair, harder to get along with, and less honest than were collaborators who did not. Interestingly, this effect was driven by one's own contribution being underappreciated more than one's collaborator's contribution being overstated. Mediational analyses revealed that the discord could be traced to the attribution of biased responsibility judgments to self-interest on the part of one's collaborator.

There is disagreement in the psychological literature over whether researchers should report one- or two-tailed tests of significance. The present research asks a slightly different question: Which approach is more compelling to research consumers? In three between-subject experiments, advanced psychology undergraduates (Experiment 1), psychology Ph.D. students (Experiment 2), and faculty at several top-ranked psychology programs (Experiment 3) read a summary of a fictitious experiment. Participants were more persuaded by the result when it was described as p