Latencies and confidence in n-value concept learning

An experiment was carried out to investigate interrelationships between responsetimes (RTs), inspection times (ITs), confidence ratings and success-error measures in concept learning tasks with 2, 4 and 6 values per dimension. On the basis of multiple hypothesis testing assumptions, ordinal predictions were made for high and low confidence RTs and ITs on error and correct trials. As predicted, high confidence RTs were significantly faster than low confidence RTs and, on correct trials, high confidence ITs were significantly faster than low confidence ITs. On error trials, high confidence ITs were (non-significantly) slower than low confidence ITs. Latency curves after the last error changed around the trial on which Ss first expressed high confidence. The latency and success-error results generally replicated previous studies.