Does Desire to Help Trump Children’s Ability to Pass a False-Belief Task?

• In Event: Theory of mind development and children’s representation of beliefs

Abstract

The cooperative bias hypothesis (Helming, Strickland & Jacob, 2016) states that children might fail false-belief tasks because they are intending to help the mistaken agent, thus turning the prediction-question of the task into a normative question. For example, in a location false-belief paradigm, instead of answering the question “Where will Sally look for X?” the participant may be really answering “Where should Sally look for X?”

In a pre-registered study, we are experimentally testing the cooperative bias by manipulating the context and urgency of the false-belief tasks. In the “neutral” condition 4- to 5-year-olds are shown a standard situation in which knowing the answer to the question has neither good or bad consequences for the agent (i.e. in an content false-belief paradigm there are paperclips instead of crayons in a crayon box); by contrast, in the “stakes” condition the agent would benefit from sharing the participant’s privileged information (i.e. there are crayons instead of band-aids in a band-aid box and the agent is hurt and needs a band-aid).

Our hypothesis is that the higher the stakes the more likely 4- to 5-year-olds will want to help the target, leading to higher failure rates. False-belief paradigms are usually tested in a between-group design as learning and carry-over effects might be an issue when focusing on individual performance. However, though individual performance in different Theory of Mind tasks might be inconsistent and noisy (Richardson, Lisandrelli, Riobueno-Naylor & Saxe, 2018; Warnell & Redcay, 2019) there are two false-belief tasks (content and location false-belief) that seem to yield a similar performance (Warnell & Redcay, 2019). In this study we wanted to see how the pragmatics of the question influence individual as well as group performance. We have been testing participants in 8 trials that combine 2 versions of each condition (2 tasks [content and location false-belief] x 2 stakes [neutral and stakes]) in a balanced Latin square design and we determined the sample size (n=72) with an a priori power analysis conducted in G*Power that gave us 80% power to detect a moderately big size effect (d=0.6) in a between-group performance.

Our research will contribute to the literature of how and when children develop the ability to pass false-belief tasks and the pragmatics that influence these cognitive processes. Moreover, this research is part of a larger project that researches intergroup theory of mind, exploring whether performance differs depending on the group membership of the agent the participant is theorizing about. The limited research on the topic seems to show that in particular tasks there might be an advantage when theorizing about the outgroup false belief which could potentially be explained by the cooperative bias hypothesis. Though this study is currently being conducted online in the US, the goal is to run a replication study in Europe. We have currently tested 50% of our total sample size and we plan complete the data collection and analysis in the next three months.

Authors

• Carlota Saumell, Universitat de Barcelona (UB)
  Presenting Author

• Mireia Hernandez, University of Barcelona
  Non-Presenting Author

• Ferran Pons, Universitat de Barcelona (UB)
  Non-Presenting Author

• Yarrow C Dunham, Yale University
  Non-Presenting Author