It was predicted that verbalization of prose materials would produce greater anxiety but would also result in greater retention of the passages, that prior knowledge of the verbalization requirement would lead to better retention, and that induced expectancy to verbalize would result in better retention than would no expectancy. Results indicated that subjects required to verbalize showed greater anxiety than did control subjects, but none of the other hypotheses was supported. A significant three-way interaction between prior knowledge, induced expectancy, and type of retention measure (cued versus recognition) was interpreted as of little practical importance in that subjects in the control conditions showed the same degree of retention as those in the verbalization conditions. Additional analyses attempting to determine the reason for the failure of the verbalization hypotheses were unable to find an adequate alternative explanation of the results. (Author/AA)