In the study of status differentiation, two dimensions of status characteristics have received special attention: the specificity-diffuseness dimension and the relevant-irrelevant dimension. A laboratory experiment manipulates a specific status characteristic, ability for "creative writing." Relevancy of the diffuse status characteristic, sex, is manipulated by having the dyad outline short story for a man's magazine, a women's magazine, or for a magazine, sex unspecified ("neutral" condition). Pre-experimental expectations have been used to explain the effects of status in small groups. In the present research, expectations are measured on two levels. As predicted, ability has an effect in same-sex dyads. Also relevancy of the task has an effect in cross-sex dyads. In mixed sex, mixed ability dyads, males given high ability feedback have more influence than females given low ability feedback. However, in the cross-sex, equal ability, neutral task condition, males do not have greater influence as predicted. Expectations are not a very satisfactory explanation for influence during dyadic interaction. Explanations for the unexpected findings are suggested as males' reaction to status threat and females' compliance to a "norm of responsibility." Suggestions are made of theoretical reformulations, practical implications, and further research. (Author/NG)