The representation of fraction magnitudes and the whole number bias reconsidered

Two experiments investigated the whole number bias in the representation of fraction magnitudes with adults. A fraction magnitude comparison task was used where half of the comparisons were consistent with whole number ordering and the other half were not. Distance effects were found in Experiment 1 indicating that participants were comparing the magnitude of the whole fraction rather than just the parts. However, accuracy and response time also depended on the comparisons' consistency with whole number ordering. Experiment 2 manipulated the distance between the fraction pairs and showed that the whole number effect was strongest when the distance between the fraction pairs was very small. The results suggest that even skilled adults do not always have direct access to a fraction's magnitude on the number line. When the magnitudes are especially close together, adults may rely on alternative implicit or explicit strategies, such as examining the whole number parts, to evaluate the comparison. (C) 2014 Published by Elsevier Ltd.