Using logistic regression to estimate delay-discounting functions.

摘要: The monetary choice questionnaire (MCQ) and similar computer tasks ask preference questions in order to ascertain indifference, the perceived equivalence of immediate versus larger delayed rewards. Indifference data are then fitted with a hyperbolic function, summarizing decline value delay time. We present fitting method that estimates parameterk directly from survey responses. Binary preferences modeled as function time (X2) transformed reward ratio (X 1), yielding logistic regression coefficients β2 β1. parameter emerges = β2/β1, where predictedp .5 (the definition indifference). MCQ was administered 1,073 adolescents scored using both standard methods. means for ln(k) were (standard, −4.53; logistic, −4.51), results highly correlated (ρ .973). Simulated showed thatk unbiased, except β1 ≥ −1, indicating vague response. Jackknife errors provided excellent coverage.