Superior exam performance for lectures with instructor self-disclosure might then be due to the decrease in memory load. Since less material was covered, there was less to be recalled. This explanation, however, is unlikely, given the overall pattern of the results. Additional class time was also required for pair-share experiences. A memory load explanation would predict that lectures that included both pair-share and instructor self-disclosure would not be able to cover as many facts and would have the smallest memory load. Yet exam performance did not reflect an advantage for this lecture condition.

I did not find support for the hypothesis concerning the positive effects of self-reference on recall of lecture material. Contrary to the findings of Dorr, Dill, Anderson, and Heppner (1996), exam performance was better on questions from lectures without student pair-share activities, than on questions from lectures with pair-share. Pair-Share decreased rather than aided recall. This result was both a surprise and a disappointment. Hartlep and Forsyth (2000) found self-reference improved students' retention of textbook material. I was searching for ways to get students to make self-reference connections more often and in different learning environments, and pair-share seemed a relatively efficient technique for classroom use. However, the result was the reverse of what I expected.