Repetition, order of presentation, and timing of arguments and measures as determinants of opinion change.

A model of order effects developed by Miller and Campbell (1959) predicts more recency with intervals between arguments and less recency with a delay of the posttest. To test the model, this study used 320 undergraduates as Ss, periods of 1 wk. as intervals and delays, and jury-trial excerpts as materials. Retention data supported the model completely. Opinion data showed less recency with delay (p<.01), as predicted, but did not show more recency with intervals—a striking nonconfirmation. Repeating an argument (1 vs. 3 times) increased opinion change slightly (p < .OS) and retention markedly (p < .01). Both retention and opinion showed strong overall recency (p < .01).