> natural experiment on 30 regular gambles, compared to control group of 30 occasional gamblers - played on a fruit machine whilst verbalising their thoughts, were interviewed afterwards
> regular gamblers saw themselves as skilful at the fruit machine, made more irrational statements, are were more likely to explain losses as 'near wins'
> supports role of cognitive biases
discussion - 'thinking aloud' research has been questioned as not representing what someone truly believes