A study examined an interaction effect observed in an earlier study in which texts about three obscure religions were presented to undergraduates who studied them under three conditions: with a study guide that analogized the religion to Christianity; with a study guide that did not employ analogies; and without aid. Scores were significantly lower among subjects using a study guide employing analogies, but since the passage to be learned contained words that explicitly signaled the same analogy as in the study guide, subjects may have confused elements of the two. In the present study, 87 undergraduate students randomly divided into two groups studied two passages with the aid of study guides. One group first read a text designed to serve as an analog that was similar to the target text, while the other group first read a text designed to serve as an analog that was in contrast to the target text. Subjects then filled out study guides about the target text while looking back on the analog. Subjects also responded to a dependent measure which asked them to list facts from memory about the target text. Results indicated that: (1) for the contrasting analogy condition, subjects wrote more correct statements than for the similar condition; and (2) for the similar analogy condition, subjects wrote more incorrect statements than for the contrasting condition, and subjects also wrote more incorrect statements that directly referred to the analog than for the contrasting condition. Findings suggest that the interaction in the earlier experiment resulted from conceptual interference set up by similarities between elements of the analogies in the study guide and the text. (Two tables of data are included; 24 references are attached.) (RS)