Examining the range of normal intraindividual variability in neuropsychological test performance

David J. Schretlen, Cynthia A. Munro, James C. Anthony, Godfrey D. Pearlson

Research output: Contribution to journalArticlepeer-review


Neuropsychologists often diagnose cerebral dysfunction based, in part, on marked variation in an individual's cognitive test performance. However, little is known about what constitutes the normal range of intraindividual variation. In this study, after excluding 54 individuals with significant health problems, we derived 32 z-transformed scores from 15 tests administered to 197 adult participants in a study of normal aging. The difference between each person's highest and lowest scores was computed to assess his or her maximum discrepancy (MD). The resulting MD values ranged from 1.6 to 6.1 meaning that the smallest MD shown by any person was 1.6 standard deviations (SDs) and the largest MD shown by any person was 6.1 SDs. Sixty-six percent of participants produced MD values that exceeded 3 SDs. Eliminating each person's highest and lowest test scores decreased their MDs, but 27% of the participants still produced MD values exceeding 3. Although MD values appeared to increase with age, adjusting test scores for age, which is standard in clinical practice, did not correct for this. These data reveal that marked intraindividual variability is very common in normal adults, and underscore the need to base diagnostic inferences on clinically recognizable patterns rather than psychometric variability alone.

Original languageEnglish (US)
Pages (from-to)864-870
Number of pages7
JournalJournal of the International Neuropsychological Society
Issue number6
StatePublished - Sep 2003


  • Clinical inference
  • Intraindividual variability
  • Neuropsychology

ASJC Scopus subject areas

  • Neuroscience(all)
  • Clinical Psychology
  • Clinical Neurology
  • Psychiatry and Mental health


Dive into the research topics of 'Examining the range of normal intraindividual variability in neuropsychological test performance'. Together they form a unique fingerprint.

Cite this