Social preferences and social influence effects ("peer effects") are well documented, but little is known about how peers shape social preferences. Settings where social preferences matter are often situations where peer effects are likely too. In a gift-exchange experiment with independent payoffs between two agents we find causal evidence for peer effects. Efforts are positively correlated but with a kink: agents follow a low-performing but not a high-performing peer. This contradicts major theories of social preferences which predict that efforts are unrelated, or negatively related. Some theories allow for positively-related efforts but cannot explain most observations. Conformism, norm following and social esteem are candidate explanations.