publications

2026

1. Interrogating Early Word Knowledge: Factors That Influence the Alignment Between Caregiver-Report and Experimental Measures

   Haley Weaver and Jenny R. Saffran

   Developmental Science, 2026

   Abs DOI

   ABSTRACT Questions about early word knowledge pervade the literature on both typical and atypical language trajectories. To determine which words an infant knows, researchers have relied on two types of measures—caregiver-report and eye-gaze behavior. When these measures are compared, however, their results frequently fail to converge, making it difficult to ascertain whether a given infant knows a given set of words. What are the reasons for these misalignments in gold-standard tasks that are designed to investigate the same underlying construct, and can convergent validity be improved? The current study was designed to investigate multiple methodological features of caregiver-report and looking-while-listening (LWL) tasks hypothesized to contribute to their alignment. American English-learning infants (18–20 months; N = 52) completed an LWL task assessing their understanding of eight early-acquired words. Caregivers reported their child’s knowledge of the same eight words, as well as their confidence in their responses and the amount of time they spend with their child. Overall, caregivers’ reports of word knowledge did not predict infants’ eye-gaze behavior. However, the measures were more likely to be aligned when caregivers reported higher confidence in their responses. Caregivers’ reports about both the target and the distractor word on each trial were related to infants’ eye-gaze behavior, suggesting that LWL tasks capture knowledge about the labels of both objects tested, not just the label of the target object. The results suggest several critical methodological modifications that could be implemented to improve the measurement validity of both caregiver-report and eye-gaze measures of word comprehension. Summary This study provides novel insights to improve the validity of infant vocabulary measurement by highlighting factors that explain differences between caregivers’ reports and eye-gaze behavior. Caregivers’ reports of individual word knowledge failed to converge with infants’ eye-gaze behavior during a looking-while-listening (LWL) task. Caregivers’ reports of individual word knowledge aligned more strongly with their infant’s eye-gaze behavior when caregivers reported more confidence in their responses. Caregivers’ reports of both the target and the distractor words predicted infants’ eye gaze behavior.

2025

1. Object functions and words reexamined: toddlers’ recognition of function depends on object type

   Haley Weaver and Saffran Saffran

   Journal of Cognition and Development, 2025

   Abs DOI

   The role of function in toddlers’ object labeling has been debated for decades in developmental science. We aimed to clarify the relation between toddlers’ understanding of functions and words using a set of everyday objects that varied in the number of associated functions (e.g., balls can be bounced, thrown, or rolled while toothbrushes primarily brush teeth). Forty 23- to 25-month-old monolingual English-learning toddlers in the U.S. completed a preferential looking paradigm in which objects were used in conventional and unconventional ways, designed to measure expectations about object functional expectations. We also measured toddlers’ lexical knowledge about these objects using a looking-while-listening task. Finally, we assessed productive vocabulary size using the MacArthur-Bates Communicative Development Inventory: Words and Sentences. The results suggest that toddlers have expectations about the functions of some objects, but not others. In particular, these expectations were stronger for objects that are tightly linked with their functions in everyday experiences, and for children who have larger vocabularies. These findings also suggest that toddlers’ ability to demonstrate functional knowledge may depend on the specific objects included in the task.

2. For preschoolers, word knowledge falls on a continuum: A novel framework for capturing the incremental process of word learning

   Rebecca A. Dore, Molly Scott, Haley Weaver, and 8 more authors

   Journal of Child Language, 2025

   Abs DOI

   In gaining word knowledge, children’s semantic representations are initially imprecise before becoming gradually refined. We developed and tested a framework for a digital receptive vocabulary assessment that captured varied levels of representation as children learn words. At pre-test and post-test, children selected one of four images to match a word’s meaning: a correct target, a conceptually-related foil, a thematically-related foil, and a phonologically-similar foil. We expected that selecting a conceptually related foil would indicate that the word is understood at a deeper level than selecting a phonologically similar foil. Indeed, selection of phonological foils decreased from pre- to post-test, while selection of more advanced thematic and conceptual foils increased. These results demonstrate that this assessment tool probed semantic knowledge that might be characterized as intermediate word knowledge. The current paper presents a novel and sensitive way to capture the incremental process of word learning. Applications for vocabulary interventions are discussed.

2024

1. Becoming word meaning experts: Infants’ processing of familiar words in the context of typical and atypical exemplars

   Haley Weaver, Martin Zettersten, and Jenny R. Saffran

   Child Development, 2024

   Abs DOI

   Abstract How do infants become word meaning experts? This registered report investigated the structure of infants’ early lexical representations by manipulating the typicality of exemplars from familiar animal categories. 14- to 18-month-old infants (N = 84; 51 female; M = 15.7 months; race/ethnicity: 64% White, 8% Asian, 2% Hispanic, 1% Black, and 23% multiple categories; participating 2022–2023) were tested on their ability to recognize typical and atypical category exemplars after hearing familiar basic-level category labels. Infants robustly recognized both typical (d = 0.79, 95% CI [0.54, 1.03]) and atypical (d = 0.70, 95% CI [0.46, 0.94]) exemplars, with no significant difference between typicality conditions (d = 0.14, 95% CI [−0.08, 0.35]). These results support a broad-to-narrow account of infants’ early word meanings. Implications for the role of experience in the development of lexical knowledge are discussed.

2022

1. Zero in on This: Children are Exposed to Various Concepts of "Zero" Prior to Age Six

   Nicholas Vest, Haley Weaver, and Martha W. Alibali

   In Proceedings of the 44th Annual Meeting of the Cognitive Science Society, CogSci 2022, Toronto, ON, Canada, July 27-30, 2022, 2022

2021

1. A framework for online experimenter-moderated looking-time studies assessing infants’ linguistic knowledge

   Desia Bacon, Haley Weaver, and Jenny Saffran

   Frontiers in Psychology, 2021

   Abs DOI

   Online data collection methods pose unique challenges and opportunities for infant researchers. Looking-time measures require relative timing precision to link eye-gaze behavior to stimulus presentation, particularly for tasks that require visual stimuli to be temporally linked to auditory stimuli, which may be disrupted when studies are delivered online. Concurrently, by widening potential geographic recruitment areas, online data collection may also provide an opportunity to diversify participant samples that are not possible given in-lab data collection. To date, there is limited information about these potential challenges and opportunities. In Study 1, twenty-one 23- to 26-month-olds participated in an experimenter-moderated looking-time paradigm that was administered via the video conferencing platform Zoom, attempting to recreate in-lab data collection using a looking-while-listening paradigm. Data collected virtually approximated results from in-lab samples of familiar word recognition, after minimal corrections to account for timing variability. We also found that the procedures were robust to a wide range of internet speeds, increasing the range of potential participants. However, despite the use of an online task, the participants in Study 1 were demographically unrepresentative, as typically observed with in-person studies in our geographic area. The potentially wider reach of online data collection methods presents an opportunity to recruit larger, more representative samples than those traditionally found in lab-based infant research, which is crucial for conducting generalizable human-subjects research. In Study 2, microtargeted Facebook advertisements for online studies were directed at two geographic locations that are comparable in population size but vary widely in demographic and socioeconomic factors. We successfully elicited sign-up responses from caregivers in neighborhoods that are far more diverse than the local University community in which we conduct our in-person studies. The current studies provide a framework for infancy researchers to conduct remote eye-gaze studies by identifying best practices for recruitment, design, and analysis. Moderated online data collection can provide considerable benefits to the diversification of infant research, with minimal impact on the timing precision and usability of the resultant data.