We use survey data from all accepted applicants to Teach for America 1993-98 to assess the longer-term effect of youth service on participants' current civic attitudes and behaviors. While TFA "graduates" score higher than the two comparison groups--"drop-outs" and "non-matriculants"--on a broad range of attitudinal items measuring civic commitment, these differences appear to be less a byproduct of the TFA experience than a reflection of current involvement with the TFA organization. Moreover, the attitudinal differences are not reflected in actual civic behavior. Specifically, graduates lag behind non-matriculants in current service activity and generally trail both non-matriculants and drop-outs in self-reported participation in five other forms of civic/political activity measured in the study. Graduates also vote at lower rates than the other two groups. Finally, fewer graduates report employment in "pro social" jobs than either non-matriculants or drop-outs. We close by speculating on what mechanisms may help explain variation in the long-term effects of youth service or activist experiences. (Contains 2 notes and 8 tables.)