What is the test-retest reliability for CNS Vital Signs?
Test-retest reliability was evaluated in a sample of 99 Ss who took the entire battery on two separate occasions, separated, on the average, by 62 days. All of the reliability coefficients are significant (p, <.05) and are comparable to or better than those reported for similar tests. The lowest correlations coefficients were obtained on the CPT variables. This, unfortunately, is not a problem with the CNSVS battery, but is something that is inherent to the CPT reports. The problem with reliability on the CPT is that normal people make few if any errors, and people with attention disorders are, by the very nature of their problem, extremely variable from test to test. Test scores are therefore not normally distributed. Kendalls tau correlation coefficients were calculated for the CNSVS CPT, which are more accurate for variables that are not normally distributed. (Pearsons r is calculated for the other tests.) If one simply scores the CPT as normal or abnormal based on a cutoff score, a
