Regarding evidence to support a link between these H-H values and possible cheating, the SCheck
program, operating with its default settings, suggested that the response-independence hypothesis could be rejected only for the student pair with H-H=7.50. Were these two students seated next to each other? Yes, they were. They both had the same family name.
I next asked the SCheck program to relax its control for false positives somewhat6; after this it rejected the response-independence hypothesis for a total of four student pairs: those shown in Table 2 with H-H values of 7.50, 1.21, and 1.14 (with EEIC=16), and one pair, not shown in the table, with an H-H index of 1.00 (EEIC of 17, D of 17). Of these four pairs, only one sat the test at the same venue, that with H-H=7.50, implying that relaxing SCheck’s Type I error rate control has produced three false positives.