Scaling of (advanced) ToM and modeling children’s ToM competencies: Longitudinal findings in 4- to 8-year-olds

• In Event: 2-136 - Social cognition beyond preschool: Advances and issues in measuring advanced theory of mind

Integrative Statement

First order theory of mind (ToM) development has shown to conform to a Guttman scale, with the understanding of desires developing before the understanding of beliefs (Wellman & Liu, 2004). There have been attempts to test for internal consistency and scalability in advanced ToM, but not over a broad age range (from first order to advanced ToM) and with a limited set of tasks. In the present study, we test for the scalability of a broader range of first order ToM tasks than were included in the classic Wellman scales, and we attempt to model the developmental transition from first order to advanced ToM in the age range from 4 to 8 years. We also include a measure of metacognition of own ignorance, based on a definition of ToM as a conceptual framework encompassing mental state ascriptions to both self and others.

We followed 155 children from the ages of 4 to 8. First order ToM was assessed with two false-belief tasks (unexpected content and unexpected location; Wellman & Liu, 2004), two morally-relevant false-belief tasks (accidental transgressor; Killen, Mulvey, Richardson, Jampol, & Woodward, 2011), and with tasks that assess children’s understanding of mental verbs (Moore, Bryant, & Furrow, 1989), their emotion understanding (Ensor & Hughes, 2005), and their metacognitive awareness of their own ignorance (Rohwer, Kloo, & Perner, 2012). The measures of advanced ToM included assessments of higher-order false belief (Coull, Leekam, & Bennett, 2006) and of children’s understanding of nonliteral speech, such as lies, white lies, metaphors, or irony (strange stories; Happé, 1994). Table 1 gives an overview of the tasks that were used at each of the four times of measurement (at the ages 4, 5, 6, and 8).

Rasch analyses showed that advanced ToM was more difficult than first order ToM. Within first order ToM (including morally-relevant false belief, mental state verbs, emotion understanding, metacognition), there was a good internal consistency, and all first order measures were Rasch scalable. Within advanced ToM, there was a low internal consistency (at age 8), and not all advanced ToM tasks could be modeled on a unitary Rasch scale together with the first order ToM tasks. This result is consistent with prior findings that indicate the involvement of multiple abilities in advanced ToM (Osterhaus, Koerber, & Sodian, 2016). Developmental conceptual coherence between first order and advanced ToM was supported by significant correlations between first order and advanced ToM (ranging from .21 to .51). A latent class analysis, however, revealed three diverse competency profiles for each of the first three times of measurement: children with competencies across a wide range of tasks; children with limited competencies on most tasks; and children with competencies on some, but not all (advanced) ToM tasks, suggesting that the currently available measures of (advanced) ToM tap different aspects of children’s developing social cognition.

Our findings support the conceptual coherence between first order and advanced ToM. At the same time, however, they suggest that the complexities involved in advanced ToM are higher than in first order ToM.

Authors

• Christopher Osterhaus, Ludwig-Maximilians-Universität München
  Presenting Author

• Daniela Kloo, Ludwig-Maximilians-Universität München
  Non-Presenting Author

• Susanne Kristen-Antonow, Ludwig-Maximilians-Universität München
  Non-Presenting Author

• Beate Sodian, LMU Munich
  Non-Presenting Author