Relations between choice and latency measures in a selective learning paradigm

Abstract Three experiments were run in a two-response selective learning paradigm to ascertain relations between (1) choice and latency on trials with both alternatives available, and (2) latency when only one alternative is available, with one alternative reinforced with greater probability. Variations used rats or pigeons, spatial or visual discrimination. One class of models of choice, accumulator models with absolute thresholds (Audley and Pike, British Journal of Mathematical and Statistical Psychology 1965, 18, 207), was examined in light of the data. Simple accumulator models, without interference between subprocesses leading to choice of the different alternatives, are clearly inadequate. An extension of Audley's accumulator model with interference (Psychological Review 1960, 67 1) makes predictions which deviate from those of the simple accumulator in the same fashion as do the data. Any attempt to account for the data by an accumulator model must provide for interference; further study of accumulators with interference would seem warranted.