Pinpointing the Cognitive Structure of the CPT-3: A Case for Distribution Appropriate Statistical Methods

Abstract

The present study used the CPT-3 as a model to illustrate the complementary use of distribution-appropriate statistical methods to analyze non-normally distributed empirical datasets. Study results reaffirmed that while group-level analysis (e.g., via traditional parametric group-level analysis or distribution-appropriate group-level analysis procedures) offers insights into performance of the group in aggregate, it is oftentimes inappropriate to presume that patterns reflected by the group are, necessarily, applicable to a smaller subset of respondents. Thus, understanding how subgroups within the population navigate and approach a given task can have direct implications for more personalized/individualized assessment and treatment, especially in clinical and research contexts. Although results of the investigation are largely supportive of study hypotheses, the study is not without limitation; perhaps most notable is the smaller sample size. Thus, replicating the investigation’s significant findings with a larger dataset would prove prudent—especially as the field of neuropsychology moves toward more consistent use of non-normally-distributed computerized reaction time tasks.