ABSTRACT
The method of successive intervals, first reported by Saffir, was designed by Thurstone to overcome some of the difficulties inherent in the method of equal-appearing intervals. The method of successive intervals is recommended when there are too many items to be scaled feasibly by the method of paired comparisons. The method of successive intervals is based on the assumption that the item judgment distributions are normal on the derived scale. Since the scale factor and zero point are arbitrary, the scale factor has been set by letting all standard deviations equal unity. The methods of paired comparisons and successive intervals yield interval scales in which the metric is obtained from the variability of judgments about the items. The scales are based on variability, but it is a remarkably predictable variability. Items that are ambiguous according to this subjective criterion should be eliminated before the successive intervals scale is computed, so that the assumption of equal standard deviations becomes tenable.
