Why Adults With ADHD Weigh Risky Rewards Differently

What the research found

The choice on the screen was between two cards. One paid eighteen points with certainty. The other paid twenty points if a hidden draw went a particular way, which it would six times in ten, and nothing otherwise. Participants made twenty-four such choices, one of which the computer picked at random and paid out in cash.

A second version reversed the arithmetic. The certain card was worth six points rather than eighteen and the gamble was unchanged, which left it worth twelve points on average against a sure six. Caution was sensible in the first version and expensive in the second.

That symmetry is the point of the design, which belongs to Tycho Dekkers and Joost Agelink van Rentergem at the University of Amsterdam with Yehuda Pollak and colleagues at the Hebrew University of Jerusalem, and appeared in the Journal of Attention Disorders in 2018. On the Iowa Gambling Task and most of its relatives, the risky deck is also the losing deck, the two correlated at about −.84. A group that draws from it more often may be chasing the spread of outcomes, or may be failing to track which deck pays.

Risk seeking, as these tasks measure it, means a preference for the option whose outcomes are more widely spread. The phrase carries none of the moral weight it has in ordinary speech. Suboptimal choice means taking the option with the lower expected value, the average return if the same choice were repeated many times over.

The team first re-examined the published literature, pooling 48 effect sizes from 34 studies, with 1,144 participants with ADHD and 1,108 controls, ranging from children to adults. Where the risky option was also the lower-value one, the ADHD groups chose it more, by a standardised difference of 0.37. Where the risky option was equal in value or better, the difference was −0.03, with a confidence interval running through zero, and the moderator separating the two families of study was significant at p = .018.

Then the experiment. Eighty university students in Israel, forty with an interview-confirmed ADHD diagnosis and forty without, average age about twenty-nine and matched on sex and education, played one version or the other. Across the sample the two groups made almost the same number of gambles, and the main effect of group was flatly absent. What separated them was the interaction with condition: where the gamble paid better, the ADHD group took it less often than the controls, a difference reaching p = .048, while where it paid worse the two groups did not reliably differ. The students expected to chase the gamble hung back from it exactly when it was worth having.

What it means for you and your child

Impulsivity and an appetite for danger are usually spoken of as one disposition. In the arithmetic of this task they come apart. What the ADHD groups did, across both halves of the research, was select the option with the poorer average return, whichever side of the gamble that option happened to occupy.

Several ordinary explanations fit. Comparing two options on expected value requires holding both sets of numbers in mind while outcomes arrive to revise them, and then resisting the tempting figure on the front of the card. Working memory and inhibition do that work, and both are documented as weaker in ADHD. Simpler shortcuts are cheaper to run, and a mind that reaches for one has stopped comparing the two values at all.

The authors draw a prediction from this and mark it as one. If the difficulty lies in comparing values, then circumstances where the bolder course is the better one ought to show the same pattern inverted: the unfamiliar conversation avoided, the harder subject not attempted. None of that was tested here, and a parent watching a cautious child would be reading forward from a card game to a life.

Nor has the real-world evidence that made the question urgent gone anywhere. The elevated rates of substance use and driving accidents in ADHD populations are cited in this paper and disputed nowhere in it. The claim under test concerns the mechanism behind those outcomes, and a mechanism is not a protection.

Every participant in the experiment was an adult, recruited from a university, asked to forgo stimulant medication for twenty-four hours, and playing for roughly a quarter of a euro a point. The pooled analysis, which did not separate its findings by age, at least included studies of children and adolescents, and it is the stronger warrant for saying anything about a child at all.

The finding concerns which option gets chosen when both values are laid out on a screen and the stakes are small. Whether the same machinery governs a decision taken at speed, in a car, with friends watching, is the first limitation the authors name and the first thing their proposed next study would test. It corrects what the older tasks were able to show, and it leaves your expectations about your own child’s judgment where they were.