The central premise of Jaswal & Akhtar (J&A) is compelling, and we agree that conceptualizing autism as a lack of social motivation is misguided. We expand upon the point that, instead, autism is most likely associated with deficits in what we refer to as socio-emotional abilities, and that these deficits lead to difficulties with social interactions. By abilities, we mean maximal effort constructs, where measurement instruments assess one's best performance under optimal circumstances (Cronbach Reference Cronbach1949). This is in comparison to typical behavior constructs, where measures assess everyday behavior or preferences independent of one's ability. Unfortunately, available and popular measures of socio-emotional abilities, within the context of autism and also for typically developing adults, are plagued by psychometric problems. We highlight prominent problems in this field with a focus on measures of theory of mind, and we provide recommendations for autism research based on socio-emotional ability research with typically developing individuals.

More often than not, measures of socio-emotional abilities do poorly in terms of precision and validity. Often, elementary information concerning the quality of measures is not even reported, arguably, because the authors are unable or uninterested to communicate this essential part of the results. This is especially problematic in autism research, where effective measurement is needed for properly understanding and treating the disorder. When tests are examined, it is often clear that more psychometric work is needed. For example, researchers have found that the popular Reading the Mind in the Eyes test (Baron-Cohen et al. Reference Baron-Cohen, Wheelwright, Hill, Raste and Plumb2001) lacks a unidimensional structure. Therefore, using a singular index to express performance is inadequate. And while the test was developed to assess the first stage of theory of mind, performance is more correlated with measures assessing the ability to perceive emotion expressed in the face of others (e.g., Olderbak et al. Reference Olderbak, Wilhelm, Olaru, Geiger, Brenneman and Roberts2015) than other measures of theory of mind (Kirkland et al. Reference Kirkland, Peterson, Baker, Miller and Pulos2013). Likewise, the test is heavily confounded by verbal abilities, further muddying its construct validity (Olderbak et al. Reference Olderbak, Wilhelm, Olaru, Geiger, Brenneman and Roberts2015). Thus, psychometric research on this measure suggests the test is imprecise and lacks adequate construct validity.

Socio-emotional ability research is also limited because many tools suffer under jingle and jangle fallacies. Jingle fallacies occur when several tests are purported to measure the same thing, but in practice show little convergence with one another, implying that they assess distinct constructs (Thorndike Reference Thorndike1904). A recent meta-analysis found that the Reading the Mind in the Eyes test was only weakly correlated with two popular measures of theory of mind, the Strange Stories test (r = .29) and the Faux Pas test (r = .29; Kirkland et al. Reference Kirkland, Peterson, Baker, Miller and Pulos2013), suggesting that these tests measure different things. Jingle fallacies also occur when researchers employ maximal effort ability tests and typical behavior self-report questionnaires but use the same term for the assessed construct (e.g., emotion perception, Olderbak & Wilhelm Reference Olderbak and Wilhelm2017).

Jangle fallacies occur when tests are purported to measure different constructs, but in practice correlate so highly as to suggest that they measure the same construct (Kelley Reference Kelley1927). This fallacy is apparent in many measures of emotion perception, empathic accuracy, and theory of mind. While each construct is conceptually distinct, in practice, researchers typically use tests where participants are asked to view another person and make inferences about what the other person is thinking and/or their feelings. It is not apparent that participants need empathic accuracy or theory of mind skills to successfully complete these tests; hence, we conclude these tests most likely assess emotion perception abilities.

The implications are that many theory of mind measures lack sufficient construct validity. Instead, it is most likely that researchers are measuring emotion perception or reading comprehension. The lack of construct validity also causes the field to be insufficiently informed as to the distinct profile of impairments in socio-emotional abilities associated with autism.

J&A question prominent classic theoretical models and established phenomena of autism research, which we support as a way to develop new insights into the nature of autism. Likewise, given that research on socio-emotional abilities in autistic people is most likely limited due to poor measurement, previous findings and conclusions may need to be questioned as well. We argue that more specificity is needed in describing and accounting for autism, and distinguishing preferences, motivation, or typical behavior from effort or ability is an essential part.

The number of distinct socio-emotional abilities is unclear. There are competing theoretical models (e.g., Elfenbein & MacCann Reference Elfenbein and MacCann2017; Mayer et al. Reference Mayer, Caruso and Salovey2016), but the empirical support for each is limited (e.g., Legree et al. Reference Legree, Psotka, Robbins, Roberts, Putka and Mullins2014). However, there is strong experimental evidence that there are distinct abilities for the perception of faces; memory for faces (Wilhelm et al. Reference Wilhelm, Herzmann, Kunina, Danthiir, Schacht and Sommer2010); the perception of emotion in faces (Hildebrandt et al. Reference Hildebrandt, Sommer, Schacht and Wilhelm2015), the voice, and gestures (Schlegel et al. Reference Schlegel, Grandjean and Scherer2012); and for emotion understanding and emotion management (MacCann & Roberts Reference MacCann and Roberts2008).

We agree with J&A that new avenues in research on autism are needed. Research on typically developing adults may offer that perspective. For example, as is the case with typically developing adults, is the ability to perceive aspects of neutral faces also strongly correlated with the ability to perceive emotion in faces? Likewise, is the ability to remember emotion expressed in faces also strongly correlated with the ability to perceive emotion expressed in faces (Hildebrandt et al. Reference Hildebrandt, Sommer, Schacht and Wilhelm2015)? We also strongly encourage studies on productive socio-emotional abilities. Importantly, are there emotion expression deficits in autistic persons, and are these deficits larger than deficits in other socio-emotional abilities (Olderbak et al. Reference Olderbak, Hildebrandt, Pinkpank, Sommer and Wilhelm2014)? Finally, research should consider newly identified abilities such as emotion attention regulation (Elfenbein & MacCann Reference Elfenbein and MacCann2017) and emotion creativity (Ivcevic et al. Reference Ivcevic, Brackett and Mayer2007).

This research would improve our understanding of cognitive processing during social interactions and further add to the ideas expressed by J&A. It may also assist in better describing the symptoms of autism, distinguishing different forms of autism, and improving interventions.