Toward a complete analysis of GSR data

A critique of amplitude analyses in GSR conditioning is discussed. It was pointed out that, when amplitudes are properly analyzed with all Ss contributing equally to each repeated measures point, the concern of latency interval, definition of response probability, and distribution of GSR amplitudes are concerns which apply to analyses of magnitude, probability, and latency, as well as of amplitude. It was concluded that nothing is lost in separating magnitude into its amplitude and probability components, particularly when magnitude, as well, is included in the statistical analyses, and that, to the contrary, we stand to learn something by such a partitioning which could not be learned without it.