Measuring Emerging Number Knowledge in Toddlers

Abstract

Recent evidence suggests that infants and toddlers may recognize counting as numerically relevant long before they are able to count or understand the cardinal meaning of number words. The Give-N task, which asks children to produce sets of objects in different quantities, is commonly used to test children’s cardinal number knowledge and understanding of exact number words but does not capture children’s preliminary understanding of number words and is difficult to administer remotely. Here, we asked whether toddlers correctly map number words to the referred quantities in a two-alternative forced choice Point-to-X task (e.g., “Which has three?”). Two- to three-year-old toddlers (N = 100) completed a Give-N task and a Point-to-X task through in-person testing or online via videoconferencing software. Across number-word trials in Point-to-X, toddlers pointed to the correct image more often than predicted by chance, indicating that they had some understanding of the prompted number word that allowed them to rule out incorrect responses, despite limited understanding of exact cardinal values. No differences in Point-to-X performance were seen for children tested in-person versus remotely. Children with better understanding of exact number words as indicated on the Give-N task also answered more trials correctly in Point-to-X. Critically, in-depth analyses of Point-to-X performance for children who were identified as 1- or 2-knowers on Give-N showed that 1-knowers do not show a preliminary understanding of numbers above their knower-level, whereas 2-knowers do. As researchers move to administering assessments remotely, the Point-to-X task promises to be an easy-to-administer alternative to Give-N for measuring children’s emerging number knowledge and capturing nuances in children’s number-word knowledge that Give-N may miss.

Miscategorized subset‐knowers: Five‐ and six‐knowers can compare only the numbers they know

“…If this difference is not caused by random noise and if the hypothesis test missed the difference only due to the lack of power, then this difference should be interpreted. One possible explanation builds on the findings that children have partial knowledge of the numbers immediately beyond their known number range (Barner & Bachrach, 2010; Gunderson et al., 2015; Krajcsi & Fintor, 2023; O'Rear et al., 2020; Silver et al., 2021; K. Wagner et al., 2019). For example, a four‐knower may have some partial knowledge of the number 5.…”

Investigating associations between parent engagement and toddlers' mathematics performance

“…Children viewed a series of paired images on the computer screen and for each, were asked to point to the one (left or right side) that showed the number of objects queried (Silver et al, 2021; Appendix A). To familiarize children with the task, children were given two practice trials with two different common objects and prompted to point to one image (e.g., ‘Which has a ball?’).…”

Children’s Learning of Non-adjacent Dependencies Using a Web-Based Computer Game Setting

“…Whereas web-based data collection in adults is extensively and successfully used and several well-established RT effects have been replicated in web-based research ( Crump et al, 2013 ; Simcox and Fiez, 2014 ), there are only a few recent web-based studies with children and infants ( Scott et al, 2017 ; Nussenbaum et al, 2020 ; Rhodes et al, 2020 ; Bambha and Casasola, 2021 ; Vales et al, 2021 ). Recent studies collecting RTs with adults and children have shown little to no difference between laboratory-based and web-based samples ( de Leeuw and Motz, 2016 ; Hilbig, 2016 ; Bridges et al, 2020 ; Nussenbaum et al, 2020 ; Morini and Blair, 2021 ; Silver et al, 2021 ; Vales et al, 2021 ) as well as no big differences between browsers (e.g., Chrome and Internet Explorer) or experiment builders (e.g., Pavlovia and Gorilla) ( Kochari, 2019 ; Anwyl-Irvine et al, 2020 ; Sauter et al, 2020 ). In addition, it is difficult to clearly state whether potential differences may be any greater than the difference between two laboratory-based collected samples ( Nussenbaum et al, 2020 ).…”