Systematic administration and analysis of verbal fluency tasks: Preliminary evidence for reliable exploration of processes underlying task performance

Abstract

Verbal fluency (VF) tasks are typically scored by the number of acceptable words generated within an allotted time (i.e., total score). However, total scores do not provide insight into verbal and executive processes underlying VF task performance. Further analyses have been implemented to increase the analytical power of VF tasks, but systematic scoring guidelines are needed. We generated instructions for administration, scoring, and analyses of total scores, errors, temporal parameters, clustering, and switching with strong inter-rater reliability. To investigate the reliability of the proposed analysis, we modeled the performance of Finnish-speaking older adults (N = 50) in phonemic (/k/, /a/, and /p/) and semantic (animals) categories. Our results are in line with previous studies: We observed a higher performance on semantic than phonemic fluency (p ≤ 0.001, d = 0.91) and significant effects for education (p ≤ 0.001, d = 1.11) and gender (p ≤ 0.001, d = −1.11), but not for age (p = 0.10, d = 0.48). Most errors were repetitions. Performance declined over the allotted time frame as measured in 15-s segments (all ps < 0.001 with medium to large effect sizes). Task congruent clustering and switching were productive strategies (all ps < 0.001 with large effect sizes), and participants generated task discrepant clusters in both phonemic (p = 0.004, d = 0.69) and semantic tasks (p = 0.66, d = 0.18). The results substantiate the proposed method, providing evidence that these guidelines are a reliable starting point for VF task performance analyses in various clinical populations investigating VF task performance in depth.