We focus on an organizational situation wherein the evaluator function and ultimate buyers are distinct and when evaluators may lack information on subsequent performance of producers. In such situations as evaluators lack valid signals of a producer’s reputation, evaluators make attributions of deservingness of producers, attributions which may not be tightly coupled with subsequent performance. These attributions are influenced by the stimuli, by the type of output an actor has produced in the past. Specifically we propose that when the capabilities of an actor are: (1) related to producing new output, i.e. frontiers, then the reliance on their past capabilities increases at an increase rate than when the past output involves older vintage output (2) related to producing output that combines wide knowledge, i.e. ‘generalists’, then the reliance on their past capabilities increases at a lower rate than when the past output involves less width of knowledge. We test these predictions in the context of funding of academic scientists’ research projects using a sample of 29,859 academics funding by federal government at a large mid-western university from 1970 to 2005. The results support our hypotheses that attentional based information processing of evaluators explains why allocations are more positive with novelty of past output and less positive with generalist contrary to the assumed uniformly positive dependence on prior output under uncertainty. Furthermore our paper develops a contingent perspective of the attention based view of organizational response to heterogeneous stimuli.