A new theory -- the Proxy Model -- addresses the use of social comparison information in realistic appraisals of ability. Previous social comparison theories have not provided a systematic explanation of how social comparison information is used to answer the question: Can I do X? Recent empirical studies of social comparison have primarily attended to comparison processes motivated by self-enhancement or self-protection purposes, rather than self-assessment. The present investigation returns to the question of how comparison information is used when people want or need accurate predictions of future performance. The Proxy Model builds on insights from Heider (1958) and extends the related attributes hypothesis developed by Goethals and Darley (1977). In essence, the Proxy Model contends that future performance in some domain can be predicted by comparing with a proxy who has already completed the task of interest. An informative proxy is one who either (1) performed similarly, under conditions of maximal effort, to the comparer on a related, previous task or (2) shares related attributes with the comparer that allow him or her to infer that Proxy's maximal capability is similar to his or her own. Eight experiments are proposed; all share a common paradigm in which participants predict performance after receiving various types of comparison information. Incentive for accuracy is also manipulated, providing high vs. low motivation to make accurate estimates. The proposed studies should provide evidence for the Proxy Model, and add to deeper understanding of the psychology of self-evaluation. The proposed research both advances basic knowledge about social comparison theory and increases understanding of related phenomena for which social comparisons play a vital role, such as self-esteem, goal setting, and self-efficacy. %%% The proposed experiments are concerned with how people decide whether they are capable of achieving performance tasks involving physical or intellectual abilities. For some time, social scientists have recognized that the comparisons people make with others influence, and partially determine, our performance predictions about what we can achieve ourselves. The factors that determine which persons' performances affect us are not well understood. The proposed studies are based on a new theory, the Proxy Model, which identifies the characteristics of other people which make them useful for making performance predictions and choices for one's self. Eight experimental studies will examine basic elements and implications of the Proxy model. The proposed research will not only contribute to our understanding of a classic phenomenon in social psychology, but also provide useful information in understanding how people choose academic and job performance goals, and how social factors contribute to the maintenance of motivation to learn and perform in school and work settings. By specifying how social factors influence the predictions people make about their future academic and job performance, we also learn about the goals people are likely to strive for, and the reactions they should have when they meet or fail to meet their expectations, two topics of pressing importance for our society.