We have tried to analyze error factors such as Alternation, Repetition, Win-stay-Lose-shift, and Lose-stay-win-shift, in the single discrimination learning. Besides application of the model has no restriction on subjects because of no need of verbal reports of subjects and has no limitation to number of dimensions or values of discrimination learning, the model can be applied to maze learning or classification learning. In order to settle one of the demerits of the model, that is, taking time for analysis, application of electric computer to the analysis was tried. Two kinds of FORTRAN programs were made. One of them needs reward sequences and response sequences as data and the other needs reward sequences and outcome sequences as data.