Using the Woodcock-Johnson IV tests of cognitive abilities to detect feigned ADHD

Abstract

Research has suggested that many young adults can successfully feign ADHD, reporting clinically significant symptom levels and displaying deficits on cognitive tasks when asked to do so. Standalone performance validity tests (PVTs) have shown some success in identifying feigned ADHD, but these tests are rarely used in typical ADHD evaluation batteries. The present study attempted to develop embedded PVT indices from the Woodcock-Johnson IV Tests of Cognitive Abilities (WJ-IV). College students (N = 150) completed a battery including tasks from the WJ-IV, as well as an established standalone PVT and a rating scale measuring ADHD and related symptoms. Thirty of the students had been professionally diagnosed with ADHD; of the remaining 120 students, half were asked to perform honestly and to the best of their ability on the battery, whereas the other half were asked to try to simulate ADHD. Several processing speed and working memory scores from the WJ-IV effectively identified students feigning ADHD, detecting at least 50% of those students at score cutoffs that also maintained specificity of 90% or more, close to the efficiency of the standalone PVT. In addition, students with ADHD diagnoses generally did not show deficits on the WJ-IV. Implications for practice and future research are discussed.

Keywords:

• ADHD
• performance validity
• WJ-IV

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 The reliability of scores from time-pressured WJ-IV COG tests was estimated through test-retest studies on selected individuals in the normative sample, and the test-retest reliability coefficients were always above 0.80, although these studies were done on groups older and younger than the individuals in our sample.

2 Since many t-tests were conducted, the possibility of familywise Type-I error is raised. In all, 44 t-tests were conducted, and so the Bonferroni-corrected p-value for each statistical test would be approximately .001. However, only 1 of our 44 p-values was between .001 and .05, so the t-tests’ statistical significance can be interpreted in essentially the same way whether or not a Bonferroni correction is applied.

3 Descriptors for d values were taken from Cohen (1992).

4 Since the specificity of the MSVT is not always 100%, we chose to leave in the one Control participant who failed the MSVT. However, we also ran our analyses with that participant removed, and found that our results did not substantially change; no significant differences between the Control and Diagnosed ADHD groups became nonsignificant, or vice versa.

Additional information

Funding

This work was supported by the ‘Woodcock Institute for the Advancement of Neurocognitive Research and Applied Practice’.