ABSTRACT
Objective
The Delis-Kaplan Executive Function System (D-KEFS) is a battery of tests designed to measure executive functions (EF). Additionally, the D-KEFS contains lower-order tasks, designed to control for speed of visual scanning, sequencing, and verbal and graphomotor output. The construct and criterion validities of D-KEFS scores that are time-based are well established. However, the constructs measured by the D-KEFS error scores are poorly understood, making clinical interpretations of such scores difficult. This study examined the construct validity of D-KEFS errors committed on EF tasks and tasks designed to measure lower-order processes (i.e., non-EF tasks), across the adult lifespan.
Method
Participants were 427 adults (18–93 years) who completed the timed subtests of the D-KEFS. Four hundred two participants also completed the Push-Turn-Taptap (PTT; a separate measure of EF) to allow cross-validation.
Results
General linear regressions showed that D-KEFS errors committed on the EF tests were associated with EF timed performance (assessed using the D-KEFS time-based scores and the PTT), but only among older adults. Importantly, errors committed on the D-KEFS tasks of lower-order processes were also associated with D-KEFS time-based EF performance, and this relationship held across the adult lifespan.
Conclusions
These findings suggest that among older adults EF errors on the D-KEFS can be interpreted as indices of EF, but such interpretations are not automatically warranted for younger adults. Additionally, errors committed on non-EF tasks contained within the D-KEFS battery can be interpreted as reflecting EF weaknesses across the adult lifespan.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Notes
1. In contrast to the first four subtests, the primary scores of the remaining five subtests of the D-KEFS reflect the abilities to reason and problem solve and as such rely on the quantification of the accuracy and quality of responses.
2. This conversion affected less than 1% of the error data points.
3. SE =.122
4. SE =.243