Individual versus triadic performance on a unidimensional complementary task as a function of initial ability level.

The hier2rchical model proposed by Laughlin, Branch, and Johnson for predicting group performance on a unidimensional complementary task by an explicit consideration of the differential relevant resources of the group me~bers was tested in a multidimensional complementary task situation with group size of three. The explicit ~enents of the unidimensional model are: {a) The unique or unshared resources of the group members of equal ability levels can be combined to produce greater performance than the members of the same ability level could do either alone or with a smaller number of comparable ability members, except for the low ability members where the overlap of resources seems to be virtually complete. {b) A person of greater task relevant resources possesses all the resources of a person of less resources. (c) The shared or nonunique resources of members of comparable ability decrease as the level of relevant resources increases. From these tenents, the predicted order of improvement for each ability level was: High: H·HH (a high working with two high partners)> (H.HM = H·HL)> (H = H.MM = H·ML = H·LL); Medium: M·HH> (M·HM = M·HL)~ M.MM>M.ML >(M = M0 LL); Low: L ·HH (L ·HM = L ·HL) L ·MM> L ·ML'> L ·LL = L. The predicted order of absolute second-administration performance for the 13 conditions was: mrn > (HUM = HHL),. (H = HMM = m1L = HLL) M~lM >MML (M • MLL) LLL = L. The predicted order of performance of comparable ability triads less their controls across levels was: (HHH H)>(MMM M)> (LLL U. 231 male and 231 female college students completed the ''Moon Problem task as individuals and were trichotomized as high (H), medium (M), or low (L) ability on the basis of their scores. They then retook the problem