The authors conducted a field experiment to evaluate the effect of receiving information about the retirement savings decisions of one's peers. Non-participants and low savers in one firm's 401(k) plan received letters enabling them to enroll or increase their plan contribution rate by returning a simple reply form. Some employees were randomly assigned to receive peer information: a statement about the fraction of their coworker peers who were participating in the plan or a statement about the fraction of their coworker peers who were contributing at least 6% of their salary to the plan. Other employees were randomly assigned to receive no such peer information. They find conflicting evidence on the impact of peer information. Among nonunionized non-participants, there is some evidence that peer information leads to a small increase in participation. But among unionized non-participants, savings plan enrollment was reduced by peer information. These results highlight the possibilities and limitations of using peer information interventions to influence behavior.