This article reviews recent evaluation studies of online learning communities to provide a systematic understanding of how such communities are evaluated. Forty-two representative studies were selected and categorized into a newly developed taxonomy of online learning community evaluations. This taxonomy is divided into four components: evaluation purpose, evaluation approach, measures for evaluation, and evaluation techniques. The findings suggest that it is inappropriate to conceptualize evaluation of such communities as a one-size-fits-all, generalizable measure of "good" or "bad." Instead, we recommend a comprehensive, on-going, diagnostic approach to measuring clusters of indicators, or "syndromes", of a particular OLC and examining the causal relation assumed by the evaluators between what is measured and the success of OLC as an imputed outcome.