Kelly Kendro
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Kelly Kendro has a total of 32 co-authors. Here are the top co-authors by number of shared publications:
- Scott Jarvis
Northern Arizona University
3 shared publications Elisabeth Marchand
University of California San Diego USA
2 shared publicationsAdeli Block
1 shared publication- Alayo Tripp
University of Florida
1 shared publication Alexander Holmberg
1 shared publicationAlexander J. Holmberg
Applied Linguistics, Northern Arizona University, Flagstaff, USA
1 shared publication
Latest Publications
- Open Access
- 3 November 2026
- Published by Wiley in Developmental Science
- Vol. 30 (1)
The Development of Subitizing in Bilingual Children
What role does language‐specific experience play in the development of numerical knowledge? Previous studies have probed this question by testing bilingual learners, who have different linguistic representations of number across their two languages. Here, we take this approach to study the effect of language‐specific experience on children's ability to rapidly estimate small sets (i.e., subitizing). We tested 66 Spanish–English and German–English bilinguals, aged 3–6, and found that bilinguals made more accurate verbal estimates of small sets in their dominant number language (i.e., the language in which they could count the highest). This was despite the fact that all children in the study were able to accurately count much larger sets, and were therefore classified as “cardinal principle knowers” in both of their two languages. These results provide evidence for early emerging individual differences in estimation abilities that are due to linguistic experience. Also, they suggest that even after children have begun to learn more advanced skills (e.g., accurate counting of large sets), individual differences in earlier learned skills (e.g., subitizing of small sets) may persist. Different language‐specific experiences may therefore impose persistent effects on numerical development, with implications not only for bilingual learners, but also for monolinguals with impoverished exposure to number language. Summary: Subitizing skills differed across bilingual children's two languages, with stronger mappings in the language in which they can count the highest. Bilingual children's subitizing skills in their dominant language, but not their nondominant language, were similar to the subitizing skills of monolinguals. Cross‐linguistic differences found for small numerosities did not extend to larger sets. Disparity in knowledge of number word meanings found in bilingual subset‐knowers persisted even after children could accurately count large sets and become CP‐knowers.
- 3 November 2026
- Published by Wiley in Developmental Science
- Vol. 30 (1)
Open AccessThe Development of Subitizing in Bilingual Children
What role does language‐specific experience play in the development of numerical knowledge? Previous studies have probed this question by testing bilingual learners, who have different linguistic representations of number across their two languages. Here, we take this approach to study the effect of language‐specific experience on children's ability to rapidly estimate small sets (i.e., subitizing). We tested 66 Spanish–English and German–English bilinguals, aged 3–6, and found that bilinguals made more accurate verbal estimates of small sets in their dominant number language (i.e., the language in which they could count the highest). This was despite the fact that all children in the study were able to accurately count much larger sets, and were therefore classified as “cardinal principle knowers” in both of their two languages. These results provide evidence for early emerging individual differences in estimation abilities that are due to linguistic experience. Also, they suggest that even after children have begun to learn more advanced skills (e.g., accurate counting of large sets), individual differences in earlier learned skills (e.g., subitizing of small sets) may persist. Different language‐specific experiences may therefore impose persistent effects on numerical development, with implications not only for bilingual learners, but also for monolinguals with impoverished exposure to number language. Summary: Subitizing skills differed across bilingual children's two languages, with stronger mappings in the language in which they can count the highest. Bilingual children's subitizing skills in their dominant language, but not their nondominant language, were similar to the subitizing skills of monolinguals. Cross‐linguistic differences found for small numerosities did not extend to larger sets. Disparity in knowledge of number word meanings found in bilingual subset‐knowers persisted even after children could accurately count large sets and become CP‐knowers.
- 16 October 2026
- Published by Springer Nature in Behavior Research Methods
- Vol. 58 (11) , 324
Identifying hidden cognate effects in L2 lexical retrieval tasks: A tutorial for stimulus design
Psycholinguistic research involving multilingual participants often examines the extent to which processing of one language is related to processing of ano
- 16 October 2026
- Published by Springer Nature in Behavior Research Methods
- Vol. 58 (11) , 324
Identifying hidden cognate effects in L2 lexical retrieval tasks: A tutorial for stimulus design
Psycholinguistic research involving multilingual participants often examines the extent to which processing of one language is related to processing of ano
- Open Access
- 30 August 2026
- Published by Center for Open Science in PsyArXiv
Best practices for crosslinguistic masked priming: A tutorial using jsPsych
The masked priming paradigm (Forster & Davis, 1984) investigates whether a target item can be cued via a prime that is shown quickly to mitigate conscious awareness during processing. In the decades since the task’s inception, researchers have used this task to probe both automatic lexical retrieval processes as well as how task parameter manipulation affects priming patterns. This paper serves a dual role. First, it functions as a tutorial for implementing a masked priming task using jsPsych (De Leeuw et al., 2023), a free and open source framework for creating web-based behavioral tasks. Second, it synthesizes information from prior research about several task parameters: masking techniques, task type, outcome variables, prime–mask language directionality, stimulus-onset asynchrony, inter-stimulus intervals, prime duration, prime type, presentation duration, and visibility tests. Materials for implementing a masked priming task, including an editable .js template for use with jsPsych 8, are included in an external repository. The tutorial also demonstrates how researchers can compare expected and observed prime duration and incorporate a visibility measure to provide an additional evaluation of prime awareness. Ultimately, the goal of this paper is to help researchers make empirically informed decisions about masked priming task design and implement this task using the jsPsych ecosystem.
- 30 August 2026
- Published by Center for Open Science in PsyArXiv
Open AccessBest practices for crosslinguistic masked priming: A tutorial using jsPsych
The masked priming paradigm (Forster & Davis, 1984) investigates whether a target item can be cued via a prime that is shown quickly to mitigate conscious awareness during processing. In the decades since the task’s inception, researchers have used this task to probe both automatic lexical retrieval processes as well as how task parameter manipulation affects priming patterns. This paper serves a dual role. First, it functions as a tutorial for implementing a masked priming task using jsPsych (De Leeuw et al., 2023), a free and open source framework for creating web-based behavioral tasks. Second, it synthesizes information from prior research about several task parameters: masking techniques, task type, outcome variables, prime–mask language directionality, stimulus-onset asynchrony, inter-stimulus intervals, prime duration, prime type, presentation duration, and visibility tests. Materials for implementing a masked priming task, including an editable .js template for use with jsPsych 8, are included in an external repository. The tutorial also demonstrates how researchers can compare expected and observed prime duration and incorporate a visibility measure to provide an additional evaluation of prime awareness. Ultimately, the goal of this paper is to help researchers make empirically informed decisions about masked priming task design and implement this task using the jsPsych ecosystem.
- 2 July 2026
- Published by Elsevier in Research Methods in Applied Linguistics
- Vol. 5 (3)
Precision matters: Synthesis through terminological review in applied linguistics
- 2 July 2026
- Published by Elsevier in Research Methods in Applied Linguistics
- Vol. 5 (3)
Precision matters: Synthesis through terminological review in applied linguistics
- Open Access
- 24 April 2026
- Published by Linguistic Society of America in Proceedings of the Linguistic Society of America
- Vol. 11 (1)
Is my bilingual your bilingual? Researchers’ definitions and operationalizations of multilingual terminology
Comparisons of results from studies of “bilinguals” often assume that the participants come from a comparable “bilingual” population. This study examines consistency in researchers’ use and definition of seven terms describing an individual’s language experience: monolingual, bilingual, trilingual, multilingual, heritage language user, language learner, and language attriter. We report preliminary data from 19 language researchers who completed a four-part survey. This paper explores responses to two of the four parts: the vignette and Likert-scale items. The vignette portion asked respondents to evaluate how well each of the seven target terms fits vignette participant descriptions. The Likert-scale portion collected responses to statements about beliefs and practices. The results showed wide variation in classification of “participant” language experience in the vignettes, as well as research practices such as control group comparisons and researcher beliefs about concepts such as the critical period and trustworthiness of self-report proficiency data. Altogether, the results highlight inconsistency in use of language experience descriptors and the implications for conducting and interpreting empirical research.
- 24 April 2026
- Published by Linguistic Society of America in Proceedings of the Linguistic Society of America
- Vol. 11 (1)
Open AccessIs my bilingual your bilingual? Researchers’ definitions and operationalizations of multilingual terminology
Comparisons of results from studies of “bilinguals” often assume that the participants come from a comparable “bilingual” population. This study examines consistency in researchers’ use and definition of seven terms describing an individual’s language experience: monolingual, bilingual, trilingual, multilingual, heritage language user, language learner, and language attriter. We report preliminary data from 19 language researchers who completed a four-part survey. This paper explores responses to two of the four parts: the vignette and Likert-scale items. The vignette portion asked respondents to evaluate how well each of the seven target terms fits vignette participant descriptions. The Likert-scale portion collected responses to statements about beliefs and practices. The results showed wide variation in classification of “participant” language experience in the vignettes, as well as research practices such as control group comparisons and researcher beliefs about concepts such as the critical period and trustworthiness of self-report proficiency data. Altogether, the results highlight inconsistency in use of language experience descriptors and the implications for conducting and interpreting empirical research.
- Open Access
- 21 April 2026
- Published by Center for Open Science in PsyArXiv
Identifying hidden cognate effects in L2 lexical retrieval tasks: A tutorial for stimulus design
Psycholinguistic research involving multilingual participants often examines the extent to which processing of one language is related to processing of another, especially during semantic priming. When choosing crosslinguistic task items in L2 vocabulary research, researchers often control for dimensions such as orthographic or phonetic overlap that may lead to a task advantage that does not represent true semantic priming effects during lexical retrieval. Prior research shows that cognates may facilitate lexical retrieval across a bilingual’s languages by activating both lexical representations of a task item (e.g., Sánchez-Casas et al., 1992; Van Assche et al., 2009). However, the cognate status of words that are semantically related to a target have not been fully considered as potential confounds, even though these pairs may trigger semantic activation that similarly facilitates recognition.In this paper, we detail seven sequential steps researchers can take when evaluating crosslinguistic target pairs to avoid compromising their paradigm due to activation of cognate synonyms. This approach allows language researchers to screen stimuli for several confounding dimensions including semantic similarity, orthographic similarity, and etymological relatedness, ultimately strengthens the validity of crosslinguistic vocabulary tasks to better elicit and evaluate lexical activation in multilinguals. We conclude by showcasing our method to evaluate a sample set of English–Spanish items and provide code for researchers to evaluate their own crosslinguistic stimuli pairs.
- 21 April 2026
- Published by Center for Open Science in PsyArXiv
Open AccessIdentifying hidden cognate effects in L2 lexical retrieval tasks: A tutorial for stimulus design
Psycholinguistic research involving multilingual participants often examines the extent to which processing of one language is related to processing of another, especially during semantic priming. When choosing crosslinguistic task items in L2 vocabulary research, researchers often control for dimensions such as orthographic or phonetic overlap that may lead to a task advantage that does not represent true semantic priming effects during lexical retrieval. Prior research shows that cognates may facilitate lexical retrieval across a bilingual’s languages by activating both lexical representations of a task item (e.g., Sánchez-Casas et al., 1992; Van Assche et al., 2009). However, the cognate status of words that are semantically related to a target have not been fully considered as potential confounds, even though these pairs may trigger semantic activation that similarly facilitates recognition.In this paper, we detail seven sequential steps researchers can take when evaluating crosslinguistic target pairs to avoid compromising their paradigm due to activation of cognate synonyms. This approach allows language researchers to screen stimuli for several confounding dimensions including semantic similarity, orthographic similarity, and etymological relatedness, ultimately strengthens the validity of crosslinguistic vocabulary tasks to better elicit and evaluate lexical activation in multilinguals. We conclude by showcasing our method to evaluate a sample set of English–Spanish items and provide code for researchers to evaluate their own crosslinguistic stimuli pairs.
- Top Cited Papers
- 21 January 2026
- Published by Wiley in International Journal of Applied Linguistics
- Vol. 36 (3) , 2822-2834
Do Large Language Models Produce Texts With “Human‐Like” Lexical Diversity? Evidence From Four ChatGPT Models
The degree to which large language models produce writing that is truly human‐like remains unclear despite the extensive empirical attention that this question has received. The present study addresses this question from the perspective of lexical diversity. Specifically, the study investigates patterns of lexical diversity in texts generated by four ChatGPT models (ChatGPT‐3.5, ChatGPT‐4, ChatGPT‐o4 mini, and ChatGPT‐4.5) in comparison with texts written by L1 and L2 English participants ( n = 240) across four education levels. Six dimensions of lexical diversity were measured in each text: volume, abundance, variety‐repetition, evenness, disparity, and dispersion. Results from one‐way multivariate analyses of variance, one‐way analyses of variance, and support vector machines (SVM) revealed that the ChatGPT‐generated texts differed significantly from human‐written texts for each variable, with ChatGPT‐o4 mini and ChatGPT‐4.5 differing the most. Within these two groups, ChatGPT‐4.5 demonstrated higher levels of lexical diversity than older models despite producing fewer tokens. The human writers’ lexical diversity did not differ across subgroups (i.e., education, language status). Altogether, the results indicate that ChatGPT models do not produce human‐like responses to the target argumentative essay prompt in relation to lexical diversity, and the newer models produce less human‐like texts than older models. We discuss the implications of these results from ChatGPT for large language models in the context of language pedagogy and related applications.
- 21 January 2026
- Published by Wiley in International Journal of Applied Linguistics
- Vol. 36 (3) , 2822-2834
Top Cited PapersDo Large Language Models Produce Texts With “Human‐Like” Lexical Diversity? Evidence From Four ChatGPT Models
The degree to which large language models produce writing that is truly human‐like remains unclear despite the extensive empirical attention that this question has received. The present study addresses this question from the perspective of lexical diversity. Specifically, the study investigates patterns of lexical diversity in texts generated by four ChatGPT models (ChatGPT‐3.5, ChatGPT‐4, ChatGPT‐o4 mini, and ChatGPT‐4.5) in comparison with texts written by L1 and L2 English participants ( n = 240) across four education levels. Six dimensions of lexical diversity were measured in each text: volume, abundance, variety‐repetition, evenness, disparity, and dispersion. Results from one‐way multivariate analyses of variance, one‐way analyses of variance, and support vector machines (SVM) revealed that the ChatGPT‐generated texts differed significantly from human‐written texts for each variable, with ChatGPT‐o4 mini and ChatGPT‐4.5 differing the most. Within these two groups, ChatGPT‐4.5 demonstrated higher levels of lexical diversity than older models despite producing fewer tokens. The human writers’ lexical diversity did not differ across subgroups (i.e., education, language status). Altogether, the results indicate that ChatGPT models do not produce human‐like responses to the target argumentative essay prompt in relation to lexical diversity, and the newer models produce less human‐like texts than older models. We discuss the implications of these results from ChatGPT for large language models in the context of language pedagogy and related applications.
- Open Access
- 10 December 2025
- Published by Center for Open Science in PsyArXiv
Evaluating potential language attrition via the Supplemental Language Attrition Questionnaire
Aims and Objectives/Purpose/Research Questions: This paper introduces the Supplemental Language Attrition Questionnaire (SLAQ), a standardized instrument that evaluates whether and to what extent an individual’s previously acquired language may have undergone attrition. Current language background questionnaires identify patterns of language acquisition and use but rarely include variables thought to influence attrition. We ask whether the SLAQ shows evidence of construct and incremental validity.Design/Methodology/Approach: The SLAQ was developed through extensive review of attrition research and expert feedback. It collects self-report data about acquisition environment, peak proficiency, attitude, motivation, period of disuse, and current contact. This paper compares SLAQ responses to responses from two existing language background questionnaires and evaluates its reliability and validity.Data and Analysis: Participants completed either the LHQ (n = 36) or LEAP-Q (n = 22) and the SLAQ. Twenty-seven participants reported attrition of at least one language. Quantitative SLAQ items were analyzed using polychoric correlation matrices, item–total correlations, and descriptive statistics.Findings/Conclusions: Results indicate that existing questionnaires fail to capture attrition, with more than half of the attrited languages not mentioned in LHQ/LEAP-Q responses. Additionally, SLAQ items demonstrated minimal overlap and strong internal consistency. Thus, the SLAQ provides a comprehensive account of language experience that is tailored to attrition contexts.Originality: This is the first validated background instrument for language attrition. Its branching design and short average completion time (7.7 minutes) make the SLAQ an unobtrusive addition to multilingualism research designs.Significance/Implications: The SLAQ offers a standardized method for evaluating attrition, which contributes to cross-study comparability, improves participant classification, and supports more accurate models of multilingual experience. Its adoption has the potential to strengthen research on attrition, bilingualism, and language assessment practices.
- 10 December 2025
- Published by Center for Open Science in PsyArXiv
Open AccessEvaluating potential language attrition via the Supplemental Language Attrition Questionnaire
Aims and Objectives/Purpose/Research Questions: This paper introduces the Supplemental Language Attrition Questionnaire (SLAQ), a standardized instrument that evaluates whether and to what extent an individual’s previously acquired language may have undergone attrition. Current language background questionnaires identify patterns of language acquisition and use but rarely include variables thought to influence attrition. We ask whether the SLAQ shows evidence of construct and incremental validity.Design/Methodology/Approach: The SLAQ was developed through extensive review of attrition research and expert feedback. It collects self-report data about acquisition environment, peak proficiency, attitude, motivation, period of disuse, and current contact. This paper compares SLAQ responses to responses from two existing language background questionnaires and evaluates its reliability and validity.Data and Analysis: Participants completed either the LHQ (n = 36) or LEAP-Q (n = 22) and the SLAQ. Twenty-seven participants reported attrition of at least one language. Quantitative SLAQ items were analyzed using polychoric correlation matrices, item–total correlations, and descriptive statistics.Findings/Conclusions: Results indicate that existing questionnaires fail to capture attrition, with more than half of the attrited languages not mentioned in LHQ/LEAP-Q responses. Additionally, SLAQ items demonstrated minimal overlap and strong internal consistency. Thus, the SLAQ provides a comprehensive account of language experience that is tailored to attrition contexts.Originality: This is the first validated background instrument for language attrition. Its branching design and short average completion time (7.7 minutes) make the SLAQ an unobtrusive addition to multilingualism research designs.Significance/Implications: The SLAQ offers a standardized method for evaluating attrition, which contributes to cross-study comparability, improves participant classification, and supports more accurate models of multilingual experience. Its adoption has the potential to strengthen research on attrition, bilingualism, and language assessment practices.
- Open Access
- 4 December 2025
- Published by Center for Open Science in PsyArXiv
The Development of Subitizing in Bilingual Children
What role does language-specific experience play in the development of numerical knowledge? Previous studies have probed this question by testing bilingual learners, who have different linguistic representations of number across their two languages. Here, we take this approach to study the effect of language-specific experience on children’s ability to rapidly estimate small sets (i.e., subitizing). We tested 66 Spanish-English and German-English bilinguals, aged 3 to 6, and found that bilinguals made more accurate verbal estimates of small sets in their dominant number language (i.e., the language in which they could count the highest). This was despite the fact that all children in the study were able to accurately count much larger sets, and were therefore classified as “cardinal principle knowers” in both of their two languages. These results provide evidence for early emerging individual differences in estimation abilities that are due to linguistic experience. Also, they suggest that even after children have begun to learn more advanced skills (e.g., accurate counting of large sets), individual differences in earlier learned skills (e.g., subitizing of small sets) may persist. Different language-specific experiences may therefore impose persistent effects on numerical development, with implications not only for bilingual learners, but also for monolinguals with impoverished exposure to number language.
- 4 December 2025
- Published by Center for Open Science in PsyArXiv
Open AccessThe Development of Subitizing in Bilingual Children
What role does language-specific experience play in the development of numerical knowledge? Previous studies have probed this question by testing bilingual learners, who have different linguistic representations of number across their two languages. Here, we take this approach to study the effect of language-specific experience on children’s ability to rapidly estimate small sets (i.e., subitizing). We tested 66 Spanish-English and German-English bilinguals, aged 3 to 6, and found that bilinguals made more accurate verbal estimates of small sets in their dominant number language (i.e., the language in which they could count the highest). This was despite the fact that all children in the study were able to accurately count much larger sets, and were therefore classified as “cardinal principle knowers” in both of their two languages. These results provide evidence for early emerging individual differences in estimation abilities that are due to linguistic experience. Also, they suggest that even after children have begun to learn more advanced skills (e.g., accurate counting of large sets), individual differences in earlier learned skills (e.g., subitizing of small sets) may persist. Different language-specific experiences may therefore impose persistent effects on numerical development, with implications not only for bilingual learners, but also for monolinguals with impoverished exposure to number language.
- Open Access
- 22 December 2023
- Published by Center for Open Science in PsyArXiv
Essentialist characterizations of language are an obstacle to accuracy, progress, and justice in science
This is the first manuscript from the ROLE Collective. Please visit our website to learn more about the collective https://rolecollective.github.io/
- 22 December 2023
- Published by Center for Open Science in PsyArXiv
Open AccessEssentialist characterizations of language are an obstacle to accuracy, progress, and justice in science
This is the first manuscript from the ROLE Collective. Please visit our website to learn more about the collective https://rolecollective.github.io/