Assessing the value of choice in a token system.

Four children earned tokens for simple tasks. After each session they could trade tokens for toys.

Some days the kids picked from six toys (choice). Other days the adult picked two toys (no-choice). The team measured how hard the kids worked under each condition.

Choice did not boost work for every child. Two kids worked the same, one worked harder with choice, one worked less.

The study shows choice is not a universal reinforcer. Each child reacted differently.

Carter et al. (2020) and Frank-Crawford et al. (2018) saw the same pattern. Kids’ stated favorites did not predict how hard they would work when tasks got tough.

Hall (1992) found that equal prize quality mattered more than choice. When prizes were unequal, kids ignored choice and went for the better item.

Together these papers warn: preference rankings, whether from choice or a standard assessment, can mislead you once real work begins.

Before you add a “choice” option to your token board, test it. Run a quick concurrent-chains assessment: let the child sample choice days and no-choice days while you count responses. Keep the condition that produces the most work, even if it surprises you.