This study examined the effects of mental ability upon problem-solving processes of individuals and dyads confronted by complex science problems. A 2 X 3 design was used, with two levels of grouping: individuals and dyads, and with three levels of mental ability: high (H), high plus low (HL), and low (L). The experimental problem was to solve the riddle of the frozen wooly mammoth, which required the subject to hypothesize, request information, and form new hypotheses. An interactive computer assisted instruction system was used as a data source, feedback mechanism, and monitoring system. Eighty undergraduates participated. The dependent variables were: number of hypotheses; number of data requests; number of data matches; total times; time expended for hypothesis; and proportion of positive, negative, and neutral information. A between-subjects ANOVA was used to analyze data. Results showed that H dyads requested significantly more data than H individuals, while L dyads received significantly fewer data matches than L individuals and HL dyads. H individuals spent significantly less time in the problem space than H dyads, and L individuals spent significantly more time than dyads. L subjects received significantly fewer instances of positive information, more neutral information, and spent more time reading neutral information than HL subjects of H subjects. H subjects received significantly more positive information, less neutral information, and spent less time reading the neutral information. (Author/DT)