Young children's difficulty on the Children's Gambling Task: Complexity or Variability of Losses?

Abstract

Gambling involves evaluating the gains and losses associated with particular actions. We investigated the extent to which 3-, 4- and 5-year-old children are able to integrate gain and loss values in order to maximize their winnings when selecting cards from two decks. This research employed the Children Gambling Task (CGT), which is a modified 2-deck version of the Iowa Gambling task used with adults. In the CGT, one deck is disadvantageous. Cards in this deck yield high gains, but higher losses and an overall net loss. The other deck is advantageous. Cards in this deck yield low gains, but even lower losses, and an overall net gain. Published research suggests that young children's difficulty with the CGT is due to task complexity. It shows that 3-, 4-, and 5-year-olds choose cards from the advantageous deck on less complex versions of the task that require them to consider losses alone (with gains held constant), or to consider gains alone (with losses held constant). However on the standard version of the CGT in which gains and losses must be considered in the same decision, 3-year-olds continue to select from the disadvantageous deck. The 5-year-olds select from the advantageous deck. The performance of 4-year-olds is more variable being above chance level in some but not all studies. An alternative interpretation of 3-year-olds' difficulty is plausible. Loss values in the disadvantageous deck of the standard CGT were more variable (0, 4, 5 or 6) than in the less complex versions (0 or 1) or (0, 5). Therefore 3-year-olds' difficulty might be attributable to variability rather than complexity. Our experiment involved 46, 3- to 5-year-old children. The variability of loss values was equated across the three versions. The findings show that (i) the complexity effect for 3-year-olds was not eliminated, and (ii) an independent measure of the ability to process complex relations significantly predicted performance on the CGT. Thus variability of loss values does not fully account for previous findings, and the complexity explanations were supported.