Decision-making skills in ASD: performance on the Iowa Gambling Task.

Ohan et al. (2015) asked adults with and without autism to play the Iowa Gambling Task. The task looks like a card game. Players pick from four decks. Some decks give big wins then bigger losses. Other decks give small wins but smaller losses. Good players learn to stick with the safer decks.

The team tracked how often each person switched decks and how fast they learned the best strategy.

The autism group picked more cards from the risky decks. They also changed decks more often. This extra switching slowed their learning. In the end they earned less play money than the control group.

The study says the problem is not just picking badly. It is changing choices too much, so the brain never locks in the better pattern.

Leung et al. (2014) looked at every lab test of "cognitive flexibility" in autism. They found most tests, like the IGT, do not clearly separate autistic from typical people. Only a parent-report checklist called the BRIEF Shift scale did. So the poor IGT scores here are real, but they are not a sure-fire diagnostic tool.

Bergmann et al. (2019) took the next step. They trained 5- to 7-year-olds with autism to shift sets in a computer game. After a few weeks the kids became more flexible and showed fewer repetitive behaviors. This suggests the deck-switching weakness L et al. found can be improved if you practice shifting in a structured way.

Alderson-Day (2011) saw a similar planning problem. Autistic kids could think of good questions in a Twenty Questions game but kept hopping to new topics. Like the IGT players, they struggled to stick with a winning plan.

If your client hops between tasks or reinforcers, they may miss the payoff, just like the gamblers. Start with fewer choices and give clear cues when a choice is "good." Model staying put. You can also borrow the SSIT game idea: short, fun shifting drills that reward sticking with a rule before changing to a new one. Over time this may cut repetitive switching and help smarter real-life decisions.