Miller and Rohling (2001) proposed a 24-step algorithm, the Rohling Interpretive Method (RIM), for quantitative interpretation of results from flexible neuropsychological test batteries. We believe that the RIM as presented in that paper has several conceptual problems, including (a) a failure to distinguish "statistically significant" from pathological differences, (b) an assumption that declines in specific abilities can be inferred when a particular test score deviates from an estimate of general premorbid ability, and (c) confusion between the standard deviation associated with individual test scores versus that of a composite of those scores. As an alternative, we suggest the value of developing and using co-normed comprehensive neuropsychological test batteries from which test users might select subsets of tests.