Observed effects of “distributional learning” may not relate to the number of peaks. A test of “dispersion” as a confounding factor