Data that have been collected for specific purposes may be exploited in the service of other interests (such as institutional research), though the exploitation may have to be accompanied by methodological compromises of various kinds. As a consequence, the findings might satisfy the criterion of being ‘good enough’ for the purposes of practical action while falling short of the expectations that would apply to original research in the social sciences. A number of examples related to the secondary analysis of datasets from higher education are presented to illustrate some of the challenges faced by analysts of existing datasets. This article stresses the need to examine given data carefully, to mitigate weaknesses in the data, where feasible, to indicate the limitations inherent in secondary analysis and to present the outcomes of analyses in a manner appropriate to the putative audience.