Four experiments explored effects on analogical transfer of evaluating solutions to base problems. In contrast to reports of positive effects of explanation, evaluation consistently reduced transfer rates and impaired mental representations of base material. This effect was not ameliorated by encoding for a later memory test, summarizing, or engaging in similar processes at encoding and recall. However, providing a prior explanation task removed the inhibitory effect of evaluation. It appears that evaluation leads to encoding of extraneous material that interferes with access to solution-critical analogous information. Prior explanation inoculates against negative effects on transfer by ensuring that new information introduced via evaluation is organized around existing representations of relevant information of the base problem. The results suggest that the source of difficulty in analogical transfer may reside not only in retrieval and mapping but also in the initial encoding of problems.