Investigating Adults’ Strategy Use During Proportional Comparison
Adults show numerical interference during discrete proportional reasoning. Although children’s similar errors are attributed to incorrect counting strategies, it is unlikely that adults use a counting strategy. We investigate two behavioral phenomena of proportional reasoning, numerical interference errors and holistic ratio-dependent responding, and use a Bayesian model-based approach to test whether these behavioral patterns can be explained by adults’ differential use of numerator comparison versus proportion comparison strategies. We find evidence of numerator interference and holistic ratio dependent responding for both discrete (i.e., individual dots) and continuous (i.e., undivided pie charts) proportions, but numerical interference is stronger for discrete stimuli. Importantly, adults’ continuous proportion comparisons were best captured by a proportion strategy, whereas discrete proportion comparisons showed a mixed pattern, with a slight preference for a numerator strategy. These findings provide insight into the mechanisms underlying proportional reasoning and provide a novel model-based approach for investigating strategy use.