Description:
<jats:p> Abstract. Aims: Longitudinal and intervention studies are needed on impaired decision-making and cognitive control deficits as putative risk factors for Substance-Related and Addictive Disorders (SAD). Adequate stability of measures is essential for this approach. To improve our knowledge, we aimed 1) to analyse retest reliabilities of such behavioural measures and 2) to compare retest reliabilities between SAD and controls. Methods: In a quasi-experimental design we recruited a convenience sample of three groups: A Gambling Disorder group (n = 26), a Nicotine Dependence group (n = 42), both diagnosed according to DSM-IV, and a healthy control group (n = 52). Participants performed two test sessions within 3 – 4 weeks with six tasks assessing decision-making and cognitive control. Results: Retest reliabilities, indicated by intraclass correlation coefficients, varied extremely between tasks and parameters ranging from 0.31 (poor) to 0.82 (excellent) with the majority ranging from 0.40 (fair) to 0.74 (good). Importantly, retest reliabilities differed significantly between the SAD groups and the control group. Conclusions: Retest reliabilities of decision-making and cognitive control measures are adequate for longitudinal and intervention studies of SAD, although tasks parameters should be selected carefully. However, group differences in retest reliabilities may result in misleading group and intervention effects. To minimize measurement error, studies investigating longitudinal designs may apply latent variable models. </jats:p>