Blanton and colleagues (1999) found that children who nominated a comparison-target in several courses chose same-sex students who slightly outperformed them in class. This had a beneficial effect on children's course grades, which were also independently predicted by comparative evaluation (i.e. how the children evaluated their relative standing in class). These phenomena were examined at two time periods with a more detailed record of comparison choices while including several psychological moderators (i.e. closeness to and identification with the comparison targets, perceived academic control, importance of academic domains). The present findings (1) replicate those found earlier by Blanton and colleagues, (2) offer evidence that children compare upward with close friends with whom they identify as a means of self-improvement, (3) show that this identification is more likely to occur when children perceive control over their standing relative to the comparison target, and (4) suggest that the effects of comparison-level choice (i.e. the level typical of the persons with whom one chooses to compare) diminish over time.