Diversity in the amount of information search and collective accuracy

Individuals differ in how much information they acquire before deciding. How does this heterogeneity affect collective accuracy when groups have the same total search resources? We examined twenty-member groups using an extension of a cue-based decision model. Agents independently inspected assigned numbers of randomly selected binary cues, integrated them with equal weights, and combined their decisions by majority vote. We enumerated all search allocations and compared those with identical total search across environments containing three, five, or ten cues. Individual search was limited to five cues. In three- and five-cue environments, the greatest feasible variation in search counts was among the best-performing allocations in every comparison. In ten-cue environments, this proportion fell to 7.7%, and homogeneous or intermediate allocations often performed better. Independent validation supported many differences, but the average and maximum accuracy losses from maximal variation were only 0.065 and 0.214 percentage points. Restricting search to odd numbers of cues reduced these differences. Search heterogeneity therefore has environment-dependent consequences even when total information acquisition is held constant. The model provides a theoretical benchmark for investigating how variation in individual information acquisition contributes to collective decisions.