Clustering of behavioral economic biases in decision-making and risk for cigarette smoking and other substance use in women and men

Background: Low loss aversion (LA) and high delay discounting (DD) are behavioral-economic decision-making biases that independently predict cigarette smoking and other risky substance use. Here we examine (1) whether low-LA and high-DD co-occur, (2) does co-occurrence increase the odds of current smoking and other substance use compared to only low-LA, high-DD, or neither; and (3) potential gender differences in these associations.

Method: Data are from five studies with U.S. adults who currently smoked or never-smoked cigarettes recruited using online convenience sampling matching on gender and education. Participants completed identical sociodemographic, substance use (cigarette, other drugs, alcohol), and LA (hypothetical 50-50 gambles) and DD (monetary-choice questionnaire) measures. LA and DD scores were dichotomized as low and high using Receiver-Operating-Characteristic Curve logistic regression.

Results: LA and DD each independently predicted substance use and with few exceptions were not influenced by gender. Low-LA compared to high-LA predicted two-fold greater odds of co-occurring high-DD (AOR = 2.120, 95%CI:1.749-2.571, p < .0001). Similarly, high-DD compared to low DD predicted two-fold greater odds of low-LA (AOR = 2.118, 95%CI:1.747-2.568, p < .0001). Among those with co-occurring low-LA and high-DD, odds of substance use were 5-10 times greater than those exhibiting neither, and 2-3 times greater than those exhibiting only low-LA or high-DD.

Conclusions: Low-LA and high-DD cluster in women and men such that exhibiting one of these decision-making biases doubles the odds of exhibiting the other. These results demonstrate reliable clustering of low-LA and high-DD and a striking increase in risk for substance use relative to having only one or neither decision-making bias.