Response format changes the reading the mind in the eyes test performance of autistic and non-autistic adults

The Reading the Mind in the Eyes Test (RMET) is a purported theory of mind measure and one that reliably differentiates autistic and non-autistic individuals. However, concerns have been raised about the validity of the measure, with some researchers suggesting that the multiple-choice format of the RMET makes it susceptible to the undue influence of compensatory strategies and verbal ability. We compared the performance of autistic (N = 70) and non-autistic (N = 71) adults on the 10-item multiple-choice RMET to that of a free-report version of the RMET. Both the autistic and non-autistic groups performed much better on the multiple-choice than the free-report RMET, suggesting that the multiple-choice format enables the use of additional strategies. Although verbal IQ was correlated with both multiple-choice and free-report RMET performance, controlling for verbal IQ did not undermine the ability of either version to discriminate autistic and non-autistic participants. Both RMET formats also demonstrated convergent validity with a well-validated adult measure of theory of mind. The multiple-choice RMET is, however, much simpler to administer and score. Lay abstract: Recognizing and understanding the perspectives of others—also called theory of mind—is important for effective communication. Studies have found that some autistic individuals have greater difficulty with theory of mind compared to non-autistic individuals. One purported theory of mind measure is the Reading the Mind in the Eyes Test (RMET). This test presents participants with photographs of pairs of eyes and asks them to identify the emotion displayed by each pair of eyes from four choices. Some researchers have argued that the multiple-choice format of the RMET may not be an accurate measure of theory of mind, as participants could simply be guessing or using a process of elimination to select the correct answer. Participants may also be disadvantaged if they are not familiar with the specific emotion words used in the multiple-choice answers. We examined whether a free-report (open-ended) format RMET would be a more valid measure of theory of mind than the multiple-choice RMET. Autistic and non-autistic adults performed better on the multiple-choice RMET than the free-report RMET. However, both versions successfully differentiated autistic and non-autistic adults, irrespective of their level of verbal ability. Performance on both versions was also correlated with another well-validated adult measure of theory of mind. Thus, the RMET’s multiple-choice format does not, of itself, appear to underpin its ability to differentiate autistic and non-autistic adults.