Test-retest reliability of a group-administered paper-pencil measure of delay discounting

The purpose of this study was to determine the reliability of an easily administered paper-pencil temporal discounting task modeled loosely on the hypothetical money choice task (Rachlin, Raineri, & Cross, 1991). Our group-administered task used 2 descriptions of discounting, the slope constant (k) of the hyperbolic function V = A/(1 + kD) and the area under the curve (AUC). The task was initially administered to 387 (163 men, 188 women, and 35 unspecified gender) introductory psychology students in groups ranging in size from 15 to 45. Six weeks later, 299 students (131 men and 168 women) repeated the task in groups ranging from 15 to 30. At each test time, the data from participants who met specified discounting criteria (about 85%) were fit well with the hyperbolic function (mean r(2) = .90). The test-retest correlation for logk was r = .64; for AUC scores, it was r = .70, with a correlation of r = -.96 between logk and AUC scores at Test 1. The (dislike for) Cognitive Complexity and Cognitive Instability subscales of the Barratt Impulsivity Scales, Version 11, administered only at the retest correlated with k values (r = .20, p = .001, and r = .179, p = .003), and women discounted more than men (p < .001) at both test times. The results are discussed in terms of the utility of the procedure and a detailed comparison of k and AUC measures of discounting.