Poster #82 - Eye Can Help: A Novel Tool for Exploring Infant Prosocial Behavior

• In Event: Poster Session 09

Abstract

Humans begin to help others early in their second year (Warneken & Tomasello, 2007) and show precursors to helping such as concern for others even earlier in infancy (Davidov et al., 2020). Studies of early helping behaviors typically rely on paradigms that require motor abilities such as grasping and handing over objects. In fact, motor abilities have been linked to toddlers’ helping (Koster et al., 2019). It is therefore difficult to determine whether young infants, who are limited in their physical and verbal abilities, would be inclined to help others if given a developmentally appropriate means of doing so. Providing infants with the opportunity to engage in helping behaviors that are less motorically taxing may provide new insight into when infants are willing to help. Because infants develop voluntary control over eye-movements before they can use their hands to intentionally manipulate objects, we developed a novel gaze-contingent paradigm for infants to manipulate objects on a screen.

Gaze-contingent eye-tracking studies have been validated with infants as young as 6 months (Wang et al., 2012) and have been used to assess infants’ social-cognitive judgements (Kanakogi et al., 2022). In the current study, we employ a gaze-contingent eye-tracking paradigm to determine 9 to 12-month old infants’ propensity to help others. Infants watch video stimuli and learn that if they fixate on buttons displayed on top of a wall for 500ms, the respective wall moves backward. Infants then watch an agent repeatedly moving to one of two target objects. In the test phase, the walls block the agent from both objects, and the infant can trigger the button on the wall blocking the goal object (helpful) or the button on the wall blocking the non-goal object (not helpful) to remove the obstacle (Figure 1).

Based on the hypothesis that infants are motivated to help others, we expect that infants will look more to the “helpful” than to the “not helpful” button. Piloting suggests infants can understand this procedure; data collection has recently begun and will conclude before the meeting. Our analysis will include a two sample t-test looking at the mean number of “helpful” button clicks compared to the mean number of “not helpful” button clicks across participants, and a one sample t-test comparing participants’ mean proportion of “helpful” button clicks (“helpful” / “helpful” + “not helpful”) to 0.50.

Authors

• Sarah Probst, University of Michigan - Ann Arbor
  Presenting Author

• Felix Warneken, University of Michigan - Ann Arbor