Abstract
An “identified point” scatter diagram may aid in program revision. Response times and error rates for program items are first obtained during laboratory study and then expressed as z scores. Pairs of z scores are plotted as points which are numbered according to the program item upon which each pair is based. The discrepancy of each point can be viewed from the location it would have occupied in the special case of perfect linear correlation. Discrepant would be the frequently incorrect, quickly answered item or the slowly but correctly answered item. These, especially, would deserve consideration in relation to the programmer's intent. Perhaps discrepancies between these or other measures of task difficulty for other tasks could also help guide task redesign.
Get full access to this article
View all access options for this article.