ABSTRACT
Young children's language and social development is influenced by the linguistic environment of their classrooms, including their interactions with teachers and peers. Measurement of the classroom linguistic environment typically relies on observational methods, often providing limited 'snapshots' of children's interactions, from which broad generalizations are made. Recent technological advances, including artificial intelligence, provide opportunities to capture children's interactions using continuous recordings representing much longer durations of time. The goal of the present study was to evaluate the accuracy of the Interaction Detection in Early Childhood Settings (IDEAS) system on 13 automated indices of language output using recordings collected from 19 children and three teachers over two weeks in an urban preschool classroom. The accuracy of language outputs processed via IDEAS were compared to ground truth via linear correlations and median absolute relative error. Findings indicate high correlations between IDEAS and ground truth data on measures of teacher and child speech, and relatively low error rates on the majority of IDEAS language output measures. Study findings indicate that IDEAS may provide a useful measurement tool for advancing knowledge about children's classroom experiences and their role in shaping development.
ABSTRACT
PURPOSE: This study examines the accuracy of Interaction Detection in Early Childhood Settings (IDEAS), a program that automatically transcribes audio files and estimates linguistic units relevant to speech-language therapy, including part-of-speech units that represent features of language complexity, such as adjectives and coordinating conjunctions. METHOD: Forty-five video-recorded speech-language therapy sessions involving 27 speech-language pathologists (SLPs) and 56 children were used. The F measure determines the accuracy of IDEAS diarization (i.e., speech segmentation and speaker classification). Two additional evaluation metrics, namely, median absolute relative error and correlation, indicate the accuracy of IDEAS for the estimation of linguistic units as compared with two conditions, namely, Oracle (manual diarization) and Voice Type Classifier (existing diarizer with acceptable accuracy). RESULTS: The high F measure for SLP talk data suggests high accuracy of IDEAS diarization for SLP talk but less so for child talk. These differences are reflected in the accuracy of IDEAS linguistic unit estimates. IDEAS median absolute relative error and correlation values for nine of the 10 SLP linguistic unit estimates meet the accuracy criteria, but none of the child linguistic unit estimates meet these criteria. The type of linguistic units also affects IDEAS accuracy. CONCLUSIONS: IDEAS was tailored to educational settings to automatically convert audio recordings into text and to provide linguistic unit estimates in speech-language therapy sessions and classroom settings. Although not perfect, IDEAS is reliable in automatically capturing and returning linguistic units, especially in SLP talk, that are relevant in research and practice. The tool offers a way to automatically measure SLP talk in clinical settings, which will support research seeking to understand how SLP talk influences children's language growth.
Subject(s)
Speech Therapy , Speech-Language Pathology , Humans , Speech-Language Pathology/methods , Child, Preschool , Speech Therapy/methods , Child , Male , Female , Speech , Video Recording , Language Therapy/methods , Linguistics , Child Language , Reproducibility of ResultsABSTRACT
PURPOSE: The purpose of this study was to examine the measurement structure of the linguistic features of speech-language pathologists' (SLPs) talk during business-as-usual therapy sessions in the public schools and to test the longitudinal stability of a theorized dimensional structure consisting of quantity, grammatical complexity, and lexical complexity. METHOD: Seventy-five SLPs' talk during therapy sessions with primary-grade students was automatically transcribed and coded for linguistic features from a corpus of 579 videotaped therapy session videos collected at the beginning, middle, and end of one school year with an approximately 12-week interval. We explored video characteristics and conducted descriptive statistics on eight linguistic indices of SLP talk to examine the variability in SLP talk between therapy sessions. Confirmatory factor analyses were used to explore the dimensional structure of SLP talk at each time point separately for the theorized three dimensions, and we conducted longitudinal measurement invariance analyses to test the stability of the three-factor structural model across the academic year. RESULTS: There were considerable variabilities among SLPs in the characteristics of SLP talk during therapy sessions. The proposed three-factor structure of SLP talk consisting of quantity, grammatical complexity, and lexical complexity had good model fit at all three time points. The linguistic measurement properties representing the three factors were invariant over time. CONCLUSIONS: Results provided robust evidence of between-SLP variability in their child-directed talk, established a three-dimensional structure of the linguistic features in SLP talk, and identified that the linguistic features in SLP talk stably measured the same constructs across one school year, based on measurement invariance. The dimensions of SLP talk during therapy with students may represent important, malleable features of therapy that influence child language gains.
Subject(s)
Communication Disorders , Speech-Language Pathology , Humans , Child , Pathologists , Speech , Child Language , Students , Speech-Language Pathology/methods , Surveys and QuestionnairesABSTRACT
This study used a short-term longitudinal design to examine the extent to which kindergartners': (a) peer relationships are associated with their school absenteeism; and (b) development of positive peer relationships vary as a function of their time spent in school. To address these aims, data were drawn from 801 kindergartners across 64 classrooms and 15 schools in a Midwest city in the United States. Results from covariate-adjusted regression models revealed that kindergartners' relational bonds were not predictive of their time away from school, and that their relationships with classmates were generally unaffected by absenteeism. There was, however, some indication that children's relationships with their classmates were more strongly predicted by their absenteeism for children from less educated households relative to children from more educated households.
ABSTRACT
Using unique data from an economically and racially diverse sample of 448 caregivers with young children (ages 4-9 years) in Ohio, we assess multiple sources of family social and economic disruptions and their associations with parenting activities during the COVID-19 stay-at-home order. Caregivers reported extensive social and economic challenges during this time, while also increasing (on average) their time spent in play/learning activities. Time spent in discipline was less likely to increase during this period. We found significant associations among disadvantaged social conditions/experiences and parenting, and that some effects were moderated by 2019 household income status. Unexpectedly, changes in economic conditions, particularly caregiver job loss, were associated with higher odds of increases in reading/telling stories time across household income groups. Overall, findings indicate that social conditions associated with the stay-at-home period of COVID-19 might have been more disruptive to parenting for caregivers with young children than the short-term economic changes.
ABSTRACT
BACKGROUND: Children with speech sound disorder (SSD) are at increased risk of reading difficulties due to poor phonological processing skills. However, the extent to which children with SSD demonstrate weaknesses on specific or all phonological processing tasks is not well understood. AIMS: To examine the phonological processing abilities of a clinically identified sample of children with SSD, with and without reading difficulties. To determine the extent to which the proportion of children with concomitant SSD and reading difficulties exhibited weaknesses in specific areas of phonological processing, or a more general phonological deficit. METHODS & PROCEDURES: Data were obtained from 157 school-aged children (Mage = 77 months, SD = 7.01) in receipt of school-based speech therapy. Approximately 25% of the sample was identified as poor readers, based on a standardized measure of word decoding. We compared the proportion of children who scored at or below 1 SD below the mean, or the bottom 16th percentile, on measures of phonological awareness, rapid automatized naming and verbal short-term memory among those identified as poor readers and good readers. OUTCOMES & RESULTS: Children with SSD demonstrated a range of phonological processing difficulties, particularly on the measure of verbal short-term memory. No specific skill differentiated groups of children with SSD with and without reading difficulties; however, those classified as poor readers on the word-decoding measure exhibited more widespread difficulties, even after controlling for language ability. CONCLUSIONS & IMPLICATIONS: Results support a cumulative risk model such that children with SSD and reading difficulties are likely to demonstrate generally poor phonological processing abilities. WHAT THIS PAPER ADDS: What is already known on the subject Children with SSD are at heightened risk of reading difficulties, particularly if their SSD persists into school age. However, not all children with SSD experience reading problems. Research aimed at determining which children are at the highest risk is mixed as to how best to identify which children with SSD are most likely to experience reading difficulties. What this paper adds to existing knowledge The study used a multiple case study approach to determine if performance on phonological processing skills might differentiate children with SSD who were poor readers from those who were good readers. As a group, children with SSD exhibited poor verbal short-term memory but relatively intact rapid automatized naming skills. No one phonological processing skill differentiated children who were poor readers from good readers. However, children with reading difficulties appeared to experience more general difficulties across phonological processing tasks, even after controlling for language abilities. What are the potential or actual clinical implications of this work? A single deficit (i.e., speech sound production) is not sufficient data to make a complete diagnosis or treatment decisions. Multiple sources of data, including several aspects of phonological processing, should be obtained to understand reading risk in children with SSD.
Subject(s)
Apraxias , Dyslexia , Language Development Disorders , Speech Sound Disorder , Stuttering , Humans , Child , Speech Sound Disorder/diagnosis , Speech Sound Disorder/therapy , Reading , Language Development Disorders/diagnosis , Language Development Disorders/therapy , Phonetics , Dyslexia/diagnosis , Dyslexia/therapyABSTRACT
Current methods for reporting interventions do not allow key questions of importance to practitioners, service providers, policy-makers and people with DLD to be answered, and hence limit the implementation of effective interventions in the real world. To extend the existing EQUATOR guidelines to the context of speech language therapy/pathology for children with language disorder and to provide more specific guidance on participants, interventions and outcomes within the CONSORT checklist (used to improve the reporting of randomised controlled trials) and TIDieR (Template for Intervention Description and Replication) to ensure consistency of reporting. We will develop a core team to include representatives from each of the key groups who will either use or be influenced by the final reporting guidance across different countries. To achieve each set of aims, we will conduct reviews of the literature (which present typologies of intervention characteristics in (D)LD and related disorders); carry out focus groups; and use systematic consensus methods such as the Delphi technique, nominal group technique or consensus development conferences. Through the development and adoption of standard intervention reporting criteria, we anticipate that we will overcome the numerous barriers for practitioners, services and policy-makers in applying intervention evidence to practice. We believe that establishing international consensus on reporting guidelines would significantly accelerate progress in DLD research and the ease with which it can be used in clinical practice, by capitalising on the growth in intervention studies to enable international collaboration and new methodologies of data pooling, meta-analyses and cross-study comparisons.
Subject(s)
Language Development Disorders , Research Design , Humans , Child , Adolescent , Consensus , Checklist , Delphi TechniqueABSTRACT
PURPOSE: The Developmental Psychobiological Model of Experiential Canalization (DPMEC) proposes that conditions of poverty-related adversity influence child self-regulation through parental caregiving, stress hormones, and the child's genetics. However, empirical findings investigating these relationships with prolonged stress hormones are mixed. Further, the relationships among conditions of adversity with prolonged stress hormones have seldom been investigated in toddlers living in poverty. Guided by the DPMEC, we examined the relationships among maternal caregiving, prolonged stress, and self-regulation in toddlers living in poverty in the United States, to include examining whether toddler prolonged stress mediated relations between maternal caregiving and child self-regulation. DESIGN AND METHODS: Participants were mothers and toddlers (20 to 24 months of age) living in poverty, who provided hair samples to measure four months of average cortisol concentration to estimate prolonged stress. We used observational measures to examine maternal caregiving and indirect report to measure children's self-regulation. RESULTS: Findings did not support the role of toddler prolonged stress in mediating the relationship between maternal caregiving and toddler self-regulation. However, multiple linear regression models showed that higher levels of maternal emotionally supportive caregiving significantly predicted better toddler soothability (b = 0.90; p = .03; 95% CI [0.10, 1.69]; partial correlation = 0.26). CONCLUSIONS: This study adds partial support for the DPMEC to represent associations between maternal caregiving and toddler self-regulation for mothers and toddlers experiencing poverty. PRACTICE IMPLICATIONS: While these data come from an observational study, pediatric nurses may consider assessing maternal supportive caregiving upon reports of poor toddler soothability.
Subject(s)
Mothers , Self-Control , Child, Preschool , Female , Humans , Hydrocortisone , Infant , Mother-Child Relations , Parents , Poverty , United StatesABSTRACT
Objective: Our study investigates how changes in family contexts were associated with child behaviors during Ohio's COVID-19 shutdown of early 2020. Background: The COVID-19 pandemic caused major economic and social changes for families. Rapid research was conducted to assess these changes and their potential impacts on child behaviors. Method: Using a diverse sample of families with children aged birth to 9 years (N = 559), we describe key economic changes and parent-reported stressors experienced during Ohio's shutdown period. Then, we use regression models to examine how these family conditions were associated with child emotional distress and changes in sleep routines. Results: When parents experienced more total COVID-19 pandemic-related stressors, they also reported that their children exhibited more anxious and withdrawn, fearful, acting out, and COVID-19 pandemic-related behaviors (p < 0.01). Conclusion: Families and children living at home in Ohio experienced significant stress during the shutdown. These findings can be used to inform future studies of the social and economic consequences of the COVID-19 pandemic for parents and children. Implications: Families and children have experienced multiple stressors during the COVID-19 pandemic. Researchers and practitioners should continue to monitor and support families and children to mitigate potential lasting consequences.
ABSTRACT
The COVID-19 pandemic and associated government-mandated shutdowns disrupted schooling, socialization, and family life for school-aged children during spring 2020. These disruptions may have contributed to increased child behavior problems. Thus, we examined behavior problems in 247 children aged 7 to 9â¯years during Ohio's shutdown period. We investigated whether differences in parent-reported child behavior problems were associated with concurrent parent distress during spring 2020 and/or children's social-emotional skills measured via teacher-reports from the previous year (spring 2019). Parent distress significantly predicted behavior problems, such that more distressed parents also reported more child behavior problems. Child pre-pandemic peer social skills also significantly predicted behavior problems, such that more skilled children exhibited fewer behavior problems. There were no interaction effects between parent distress and children's social-emotional skills on child behavior problems. Further research is needed to understand how children's social-emotional skills impact their ability to cope during times of epidemiological crisis.
ABSTRACT
Temperament is a dynamic trait that can be shaped by maturity and environmental experiences. In this study, we sought to determine whether and the extent to which temperament was predicted by sleeping behaviors in an understudied sample of primarily Black and White infants and toddlers living in low-income homes (N = 150). Sleeping behaviors were assessed at 15-19 months of age with caregiver report of the Tayside Children's Sleep Questionnaire. Temperament was examined as effortful control, negativity, and surgency with the Infant Behavior Questionnaire Very Short Form at 9-12 months of age and with the Early Childhood Behavior Questionnaire Short Form at 20-24 months of age. Covariates were maternal education, household income, and child sex and race. Continuous variables were standardized, then missing data from independent variables were multiply imputed in 20 datasets. Regression analyses showed that about 1 SD improvement in toddler sleep behaviors significantly predicted about 1/5 SD better toddler effortful control. However, sleep behaviors did not significantly predict toddler surgency or negative affect. This study shows that for a sample of infants and toddlers in low-income homes, how a child learns to regulate sleeping behaviors may influence the development of overall effortful control about six months later.
Subject(s)
Poverty , Temperament , Child , Child Behavior , Child, Preschool , Humans , Infant , Infant Behavior , SleepABSTRACT
Experiencing chronic stress early in life is associated with later health disparities, and poverty may be a significant stressor for both mothers and children. With a sample of primarily Black and White mothers (N = 75) and toddlers (N = 71) living in poverty in the United States, we examined the direct relations between sociodemographic conditions of poverty and chronic physiological stress. Mothers completed questionnaires on sociodemographics, including mother/toddler race, mother's education, father's education, poverty level, economic hardship, marital status, unemployment status, and toddler sex. Physiological chronic stress was measured by assaying the cortisol content of 4 cm samples of hair cut from the posterior vertex of mothers and toddlers (20-24 months of age) to represent 4 months of stress. Mothers' and toddlers' chronic stress was significantly, moderately, and positively associated. Toddlers had a trending relationship of moderately higher chronic stress if they were Black compared to not Black. Mothers had significantly, moderately higher chronic stress if they were Black or had a Black toddler (compared to not Black), not married (compared to married), or were working (compared to not working). The findings suggest that these mothers, simultaneously navigating poverty and parenting a toddler, need resources to reduce chronic stress.
Subject(s)
Mothers , Poverty , Child, Preschool , Female , Hair , Humans , Hydrocortisone , Infant , Mother-Child Relations , Parenting , United StatesABSTRACT
Purpose Narrative skill represents a higher-level linguistic skill that shows incremental development in the preschool years. During these years, there are considerable individual differences in this skill, with some children being highly skilled narrators (i.e., precocious) relative to peers of their age. In this study, we explored the contribution of three lower-level language skills to a range of narrative abilities, from children performing below expected levels for their age to those performing much higher than the expected levels for their age. We speculated that individual differences in lower-level skills would contribute meaningfully to variability in narrative skills. Method Using a sample of 336 children between 3 and 6 years of age (M = 4.27 years, SD = 0.65), both multiple regression and quantile regression approaches were used to explore how vocabulary, grammar, and phonological awareness account for variance in children's "narrative ability index" (NAI), an index of how children scored on the Narrative Assessment Protocol-Second Edition relative to the expected performance for their age. Results Multiple regression results indicated that lower-level language skills explained a significant amount of variance (approximately 13%) in children's NAI scores. Quantile regression results indicated that phonological awareness and vocabulary accounted for significant variance in children's NAI scores at lower quantiles. At the median quantile, vocabulary and grammar accounted for significant variance in children's NAI scores. For precocious narrators, only vocabulary accounted for a significant amount of variance in children's NAI scores. Conclusion Results indicate that lower-level language skills work in conjunction to support narrative skills at different ability levels, improving understanding of how lower-level language skills contribute across a spectrum of higher-level linguistic abilities.
Subject(s)
Child Language , Vocabulary , Awareness , Child , Child, Preschool , Humans , Linguistics , NarrationABSTRACT
Both mathematical and narrative abilities are considered critical dimensions of children's kindergarten readiness skills and are emphasized in developmental standards and readiness assessments during the prekindergarten period. Yet, despite considerable opportunities available to educators to leverage potential cross-domain associations in their instruction, little is known about how component math skills are associated with narrative skills. The current study extends prior theoretical and empirical work on cross-domain associations by examining the extent to which component math skills, constituting a kindergarten readiness assessment battery, are associated with narrative comprehension skills in children aged 4-6 years. Two geographically distinct samples of children enrolled in a summer kindergarten readiness program in the United States (N = 108) were used to explore how early numeracy skills (e.g., counting, cardinality), math language, and patterning skills individually and together contributed to variance in narrative comprehension skills. Zero-order correlations indicated that math and narrative skills were significantly and moderately correlated, with cardinality and math language skills showing the strongest associations, followed by patterning and counting skills. Furthermore, results from a multiple regression analysis indicated that cardinality and math language each explained a significant and substantial proportion of variance in narrative comprehension skills when controlling for the other math skills. Implications for theoretical models of cross-domain development of children's cognitive skills are discussed.
Subject(s)
Comprehension , Language , Child , Child Development , Child, Preschool , Humans , Mathematics , SchoolsABSTRACT
Early self-regulation is a foundation for lifelong wellness and can be shaped by the interplay among several vital exposures. In this study, we examined the emergence of reliable profiles based upon exposure to risk and protective factors in infancy, determined if sociodemographic resources predict profile membership, and determined if these profiles predict early regulatory behaviors in a sample of infants reared in low-income homes. Data were collected from a sample of primarily Black or White mother-infant dyads living in low-income homes in the Midwest (n = 222) during the infants' first year of life (mean maternal age at enrollment: 26.29 years; range 18-43 years). Exposures included mother-infant interactions; father support; maternal depression, stress, and self-efficacy; home environment; food security; and breastfeeding duration. Sociodemographic resources included poverty status; economic hardship; maternal education, employment, and age; parental marital status; and infant race and sex. Infant regulatory behaviors were measured with the Infant Behavior Questionnaire-Revised Very Short Form effortful control subscale (mean age 11 months; range 8.5-14.3 months). Latent profile analysis was used to profile infants by risk and protective exposures. Regression was used to differentiate profiles by sociodemographic resources and to predict infant regulatory behavior from profiles. Three profiles emerged: low father support, good maternal mental health, and poor maternal mental health. A married mother, less economic hardship, and working mothers predicted infant exposure to good maternal mental health. Infant regulatory behavior was best when the infant was exposed to the good maternal mental health profile. Implications for future research are discussed.
Subject(s)
Mothers , Poverty , Adolescent , Adult , Breast Feeding , Female , Humans , Infant , Mother-Child Relations , Protective Factors , Young AdultABSTRACT
The authors examined the association of maternal mental health symptoms and symptom clusters with the sleeping behaviors of toddlers living in low-income homes. The sample is comprised of 174 racially diverse mother and toddler (15-19 months old) dyads who were participating in a longitudinal birth cohort study. Dyads were initially recruited from Women, Infant, and Children clinics in a Midwestern county of the United States. Maternal mental health symptoms were measured with self-reports of depression (Center for Epidemiological Studies-Depression), parenting stress (Parenting Stress Index Very Short Form), and parenting self-efficacy. Toddler sleeping behaviors were measured with an adaptation of the Tayside Children's Sleep Questionnaire that assessed toddler difficulties in initiating or maintaining sleep (DIMS). Maternal mental health symptom clusters were identified with latent profile analysis and toddler DIMS were regressed on maternal mental health symptoms. We found that 1 SD worse maternal depression, parenting stress, or parenting self-efficacy predicted 0.18-0.24 SD worse toddler DIMS (p < .05). Moreover, we found that the cluster of adverse maternal mental health symptoms predicted nearly 0.75 SD worse toddler DIMS as compared with the cluster of good maternal mental health symptoms (p < .05). These findings provide insight into how maternal mental health symptom clusters may be important in understanding sleep behaviors in early toddlerhood, a sensitive period of sleep development, and underscore the need to support mothers with co-occurring adverse mental health symptoms, as their children may be at higher risk for DIMS.
Subject(s)
Mental Disorders/psychology , Mother-Child Relations/psychology , Mothers/psychology , Poverty/psychology , Sleep , Adult , Child Development , Child, Preschool , Cohort Studies , Cross-Sectional Studies , Female , Humans , Infant , Longitudinal Studies , Male , Maternal Behavior , Parenting , Polysomnography , Surveys and Questionnaires , United StatesABSTRACT
OBJECTIVE: This study examines the media use of children from low-income homes during school closings during the COVID-19 pandemic. METHODS: Caregivers of 151 kindergarteners from low-income homes completed questionnaires as part of a larger study. Caregivers reported how much time children spent watching television/videos and using apps on the most recent weekday and weekend days. Caregivers also reported how their child's current use of media for several different purposes compared with how much the child usually uses media for that purpose. RESULTS: Weekly average media use was 46.3 hours or 6.6 hours per day. Counter to previous research, weekday media use was higher than weekend media use, suggesting that media was likely used as a replacement for time usually spent in school. Caregivers reported increased child media use for positive purposes, such as education and maintaining relationships with family and friends outside of the home, and potentially useful but less socially valued purposes, such as occupying the child's time while caregivers were completing other tasks. Having more children in the household was related to higher media use, and girls used media for maintaining remote relationships more than boys. CONCLUSION: These findings provide reason for both concern and optimism for the impacts of pandemic closures on low-income children. High levels of media use seem to be prevalent in this population. However, the diverse purposes for media use suggest that caregivers relied on media to supplement children's academic and social growth at a time when school and socializing were not safe in their typical forms.
Subject(s)
COVID-19 , Poverty , Television , Caregivers , Child, Preschool , Family Characteristics , Female , Humans , Male , Ohio , Pandemics , Television/statistics & numerical dataABSTRACT
In the current study, we examined whether the quantity of toddlers' exposure to media was related to language skills and whether meeting the American Association of Pediatrics (AAP) recommendations of limiting media exposure to one hour or less per day was related to language skills. We examined these associations in a sample of toddlers (N = 157) living in low-income homes. Toddlers were about two years of age (M = 28.44 months, SD = 1.48 months) during the first visit when parents reported on toddlers' exposure to media in the home. Toddlers were about three years of age (M = 33.61 months, SD = 1.06 months) during the second visit when direct measures of toddlers' expressive and receptive language and receptive vocabulary skills were completed. Controls were child gender, race, mothers' education, marital status, work status, and center-based child care. Results indicated that more frequent exposure to media was related to lower expressive language, but not receptive language or receptive vocabulary. The predictor of AAP recommendation was not significantly related to any child language outcomes. These results suggest that media exposure may be related to the displacement of language-enhancing activities during a critical time for toddlers' language development. However, the AAP media recommendation of one hour may not be related to language development.
Subject(s)
Child Language , Language Development , Child , Child, Preschool , Female , Humans , Language , Mothers , VocabularyABSTRACT
Interactions with teachers and peers are critical for children's social, behavioral, and academic development in the classroom context. However, these two types of interpersonal interactions in the classroom are usually pursued via separate lines of inquiries. The current study bridges these two areas of research to examine the way in which teachers influence child-perceived peer social support and peer victimization for 2,678 children within 183 classrooms in preschool through grade three. Two levels of teacher influence are considered, namely teacher-child closeness and conflict relationships at the child-level, and teacher management of interpersonal interactions at the classroom-level. Results of multilevel regression models showed that teacher-child closeness was associated with the growth of child-perceived peer social support from fall to spring, whereas teacher-child conflict and teachers' behavior management practices were associated with the change in child-perceived peer victimization across the academic year. These associations were unique and above and beyond the influence of children's actual peer social interactions, including reciprocal friendships and the collective classroom reputation of peer victimization. Collectively, findings highlight the multi-faceted teacher roles in shaping children's perceptions of their peer social experiences during the earliest years of schooling.
ABSTRACT
The present study explored whether a tool for automatic detection and recognition of interactions and child-directed speech (CDS) in preschool classrooms could be developed, validated, and applied to non-coded video recordings representing children's classroom experiences. Using first-person video recordings collected by 13 preschool children during a morning in their classrooms, we extracted high-level audiovisual features from recordings using automatic speech recognition and computer vision services from a cloud computing provider. Using manual coding for interactions and transcriptions of CDS as reference, we trained and tested supervised classifiers and linear mappings to measure five variables of interest. We show that the supervised classifiers trained with speech activity, proximity, and high-level facial features achieve adequate accuracy in detecting interactions. Furthermore, in combination with an automatic speech recognition service, the supervised classifier achieved error rates for CDS measures that are in line with other open-source automatic decoding tools in early childhood settings. Finally, we demonstrate our tool's applicability by using it to automatically code and transcribe children's interactions and CDS exposure vertically within a classroom day (morning to afternoon) and horizontally over time (fall to winter). Developing and scaling tools for automatized capture of children's interactions with others in the preschool classroom, as well as exposure to CDS, may revolutionize scientific efforts to identify precise mechanisms that foster young children's language development.