Online measures of looking and learning in infancy

Abstract

Infants in laboratory settings look longer at impossible than possible events, learn better about objects that behave surprisingly, and match people’s utterances to the objects that likely elicited them. The paradigms that reveal these behaviors have become cornerstones of research on preverbal cognition. But less is known about whether these canonical behaviors generalize in naturalistic environments. Here we describe a series of online protocols that replicate classic laboratory findings, detailing our methods throughout. In Experiment 1, we found that 15-month-olds (N = 24) looked longer at an online impossible support event (an object appearing to defy gravity) than a possible support event. These infants did not, however, show the same success with an online solidity event. In Experiment 2, we found that 15-month-olds (N = 24) showed surprise-induced learning online—they were better able to learn a label for a novel object when the object had just behaved unexpectedly. Finally, in Experiment 3, we found that 16-month-olds (N = 20) who heard a valenced utterance (“Yum!”) showed preferential looking to the object most likely to have generated that utterance. Together, these results suggest that, with some adjustments, testing infants online is a feasible and promising approach for cognitive development research.

A Framework for Online Experimenter-Moderated Looking-Time Studies Assessing Infants’ Linguistic Knowledge

“…In addition to these platforms, developmental researchers have been using Zoom for experimenter-mediated studies. For example, Smith-Flores et al (2021) used Zoom to replicate several in-lab findings using violation of expectation paradigms with 15-month-old. Importantly, Smith-Flores et al (2021) reported global measures of looking-time (i.e., average proportion of looks), but did not examine moment-by-moment changes in visual attention in response to a stimulus, as is common with lab-based experiments focused on early language development.…”

“…For example, Smith-Flores et al (2021) used Zoom to replicate several in-lab findings using violation of expectation paradigms with 15-month-old. Importantly, Smith-Flores et al (2021) reported global measures of looking-time (i.e., average proportion of looks), but did not examine moment-by-moment changes in visual attention in response to a stimulus, as is common with lab-based experiments focused on early language development. Although Sheskin and Keil (2018) and Smith-Flores et al (2021) suggest that experimenter-moderated data collection is promising, it is unclear how variability between participants' home set-ups impacts the subsequent data quality.…”

“…Importantly, Smith-Flores et al (2021) reported global measures of looking-time (i.e., average proportion of looks), but did not examine moment-by-moment changes in visual attention in response to a stimulus, as is common with lab-based experiments focused on early language development. Although Sheskin and Keil (2018) and Smith-Flores et al (2021) suggest that experimenter-moderated data collection is promising, it is unclear how variability between participants' home set-ups impacts the subsequent data quality. Lack of internet access or poor internet connectivity could render participants' video data unusable due to inconsistent stimulus presentation.…”

Bringing Home Baby Euclid: Testing Infants’ Basic Shape Discrimination Online

“…One possible factor that may have limited infants' success in the present online experiments is the stimuli's presentation on small, personal computer screens. For example, while Smith-Flores et al (2021) found looking-time results with toddlers tested online that were largely consistent with lab-based results, they speculated that their one null-finding -that infants failed to look longer at events in which an object appeared to move through another object after rolling down a ramp -may have been due to the events' being presented on a small screen, which minimized the visibility and salience of the violating object's trajectory. Similarly, the small screens used in the present study may have limited the visual saliency of the subtle shape changes.…”

“…Several recent online studies have found results that are mostly consistent with in-lab results using either the moderated video-chat or unmoderated approach. These studies have nevertheless adapted forced-choice paradigms with children or looking-time paradigms with older infants and toddlers like preferential or "violation-of-expectation" paradigms (Leshin, RUNNING HEAD: BRINGING HOME BABY EUCLID 4 Leslie, & Rhodes, 2020;Lo et al, 2021;Nussenbaum, Scheuplein, Phaneuf, Evans, & Hartley, 2020;Sheskin & Keil, 2018;Smith-Flores, Perez, Zhang, & Feigenson, 2021).…”