Investigating retest effects in cognitive ability tests: An operation-specific approach

The term “retest effects” refers to score gains on cognitive ability as well as educational achievement tests upon repeated administration of the same or a similar test. Previous research on this phenomenon has focused mainly on general cognitive ability scores—often using manifest difference scores—and has neglected differences in retest effects across different types of cognitive operations underlying general cognitive abilities. Additionally, these studies have focused primarily on average group-level test scores, neglecting interindividual differences in retest effects. To address these gaps, we used latent growth curve modeling to examine retest effects in N = 203 participants across three test sessions, considering both general cognitive ability and its four underlying operations according to the Berlin intelligence structure model, namely, processing capacity, processing speed, creativity, and memory. Results show a linear improvement in overall performance of 53.60 points (about 10.45 IQ points) with each assessment, corresponding to two thirds of a standard deviation. Participants' slopes—that is, their rates of improvement across test sessions—did not vary significantly, and thus did not correlate with their initial cognitive ability levels. Statistically significant operation-specific differences in the magnitude of retest effects were found, with memory showing the largest retest effect and creativity the smallest. Although participants did not vary in their rates of improvement on the processing-capacity and memory operation, there was significant interindividual variation in the slopes of the other two operations. These findings highlight the importance of considering operation-specific scores in research on retest effects. Implications for cognitive ability retesting practices are discussed. • Retesting leads to linear increase in scores across three test sessions. • Retest effects vary in size across four cognitive ability operations. • Latent growth curve modeling is used to capture individual differences. • No individual variability in improvement for processing capacity and memory. • No support for matthew or compensation effect across any cognitive ability operation.