Parent-Child Shared Book Reading Mediates the Impact of Socioeconomic Status on Heritage Language Learners’ Emergent Literacy

Abstract

The promotion of parent-child shared book reading for the development of children’s literacy is a vital component of early literacy evidence-based recommendations. A robust body of research demonstrates that, regardless of socioeconomic status, parent-child shared book reading promotes monolingual children’s literacy. However, despite the growing population of heritage language learners in the United States, those who grow up in homes where a familial language is spoken but receive English instruction at school, parent-child shared book reading research among heritage language learners remains scarce. Further, the impact of parent-child shared book reading is likely to alter the influence of family’s socioeconomic status on heritage language learners’ emergent literacy development. Here, parent-child shared book reading and children’s emergent literacy were examined using latent variable path analyses in a national sample of 965 heritage language learners ages 2 to 6 years old. Parent-child shared book reading mediated the effect of socioeconomic status on heritage language learners’ emergent literacy.

The United States has been experiencing an unprecedented increase in immigration. Currently, over 60 million individuals speak a language at home that is not English (U.S. Census Bureau, 2019); this includes 12 million school-age children, accounting for 23% of the school population (Kids Count Data, 2020). This number is projected to grow dramatically in the next 50 years such that immigrants and their children will comprise about a third of the total U.S. population (Fry et al., 2015). As such, heritage language learners, individuals who grow up in homes where a non-English language is primarily spoken but receive English instruction at school (Valdés, 2001), will have a greater presence in U.S. schools. However, these children from a sociocultural minority language background have historically lagged behind academically (e.g., Callahan & Humphries, 2016). Considering the significant well-documented impact of emergent literacy on academic access (e.g., Cunningham & Stanovich, 1997; Dodici et al., 2003), evidence of how to support literacy among heritage language learners is in dire need.

The importance of parent-child shared book reading has been underscored in a plethora of studies examining its effect on children’s first language and literacy development (e.g., Debaryshe, 1993; Zivan & Horowitz-Kraus, 2020). Moreover, a small number of studies have confirmed that parent-child book reading confers similar advantages for the literacy development of bilingual children. The current study adds to the literature by exploring the relation between the frequency of parent-child book reading and the emergent literacy of heritage language learners from diverse first language backgrounds. Heritage language learners constitute a unique subset of bilinguals who maintain a continuum of heritage language abilities (i.e., from native-like to very limited proficiency; Polinsky, 2015).

Parent-Child Book Reading and Monolingual Children’s Early Literacy

Parent-child shared book reading can spark children’s interest in reading and provide a natural context for parents to help their children develop knowledge about books, print, and reading (Clay, 1979; Lewis et al., 2016). During shared book reading, a parent reads and discusses a book with a young child (van Kleeck et al., 1997). The conversations between the parent and child may expose children to language and vocabulary that they have not encountered in daily life, promote the construction of meaning in a shared context (Halliday, 2004), build children’s conceptual knowledge (Neuman et al., 2011), and provide a context to model speech-to-print connections (Justice et al., 2002).

Numerous studies on monolingual parent-child book reading have reported that shared book reading experiences contribute substantially to young children’s oral language skills, emergent literacy, and reading achievement (see Bus et al., 1995; Mol & Bus, 2011; Scarborough & Dobrich, 1994 for reviews). The literature on monolingual parent-child book reading reports that the effect of parent-child shared book reading was greater for younger children (Bus et al., 1995; Debaryshe, 1993). As such, it is not surprising that prior to formal schooling, individual differences in print exposure already vary (Baker et al., 1997; Bus, 2002). Children who are read to from an early age enter school with more oral language (Crain-Thoreson & Dale, 1999; Debaryshe, 1993), narrative skills (Harkins et al., 1994), phonological awareness (Burgess, 1997), print awareness (Justice & Ezell, 2000), larger vocabularies (Hood et al., 2008; Robbins & Ehri, 1994), and a more advanced understanding of the connections between spoken and written words (Hindman et al., 2008), which in turn leads to greater word recognition and familiarity with the spelling of words (Ziegler & Goswami, 2005). Therefore, parent-child shared book reading makes a significant difference in children’s academic lives, preparing them with the early knowledge and skills that are needed to support the process of learning to read and are known to have a long-term impact on academic success.

Parent-Child Book Reading and Bilingual Children’s Early Literacy

Schools in the United States increasingly serve a large and growing number of linguistically diverse students, many of whom struggle to develop English literacy skills at the same rate as their monolingual peers (Hussar et al., 2020). As such, there is a pressing need to understand if parent-child book reading confers the same advantages for the literacy development of children from diverse language backgrounds. Here, we briefly review the literature on parent-child book reading among bilinguals, including heritage language learners.

The importance of parent-child book reading on children’s oral language development in both their first and second language has been underscored in studies examining different components of the home literacy environment (HLE). For example, Bitetti and Hammer (2016) examined the impact of HLE on the English narrative development of Spanish-English bilingual children. These Spanish heritage language learners were followed from preschool through first grade. They found that the frequency of mother-child shared book reading was the only significant and positive predictor of the growth of the children’s total narrative scores in English. There was no significant effect of story exposure (mothers told make-believe stories), story production (children told make-believe stories), book reading, or the number of children’s books owned. However, among preschool Spanish heritage language learners, Lewis and colleagues (2016) found that the frequency of mother-child book reading was the sole significant HLE predictor of children’s Spanish vocabulary and oral comprehension but not their English vocabulary and oral comprehension.

Similar results were reported in non-heritage language learners, bilingual Singaporean kindergartners learning English as a second language from various first language backgrounds (i.e., Mandarin, Malay, Tamil). O’Brien and colleagues (2020) used a 4-factor model of HLE that consisted of parent-child book reading, child interest, parent habit, and parent involvement. Among these factors, shared book reading not only had the strongest relation with children’s English receptive vocabulary, phonological awareness, alphabet knowledge, word reading, and sentence comprehension but also significantly predicted children’s vocabulary and reading outcomes in English. Shared book reading included the frequency with which the child asked to be read to and the frequency of parent-child book reading.

A few studies have focused on whether the links between parent-child book reading and children’s literacy were language-specific; in particular, whether shared book reading facilitates children’s literacy across languages or in only the language the shared reading occurs. For instance, Patterson (2002) examined links between expressive vocabulary size and parent’s self-reported frequency of parent-child shared book reading. Expressive vocabulary refers to the words that an individual can express or produce. In longitudinal studies, the timing of expressive language development predicts later literacy development (e.g., Del Tufo et al., 2019). In both Spanish and English, the frequency that toddler Spanish heritage language learners were read to in each language was significantly associated with their expressive vocabulary size in the same language. The relation was still evident when controlling for the child’s overall exposure to each language, indicating the unique contribution of parent-child shared book reading experiences to children’s Spanish and English vocabulary. Moreover, Quiroz, Snow, and Zhao (2010) observed that for Spanish heritage language learners (4-to 5-years-old), self-reported mother-child reading frequency in each language positively predicted children’s expressive vocabulary in the same language. However, Quiroz et al. (2010) also found that mothers’ English reading negatively affected children’s Spanish expressive vocabulary. In contrast, Sun (2019) failed to identify the language-specific links between parents’ self-reported shared reading frequency in Chinese and children’s Chinese vocabulary and grammar among a group of Chinese-English bilingual preschoolers in Singapore, who are not heritage language learners. Thus, while language-specific links between parent-child shared book reading and children’s literacy were evident among Spanish-English bilinguals, there was no relation in Chinese-English bilinguals’ Chinese.

Studies using print exposure checklists reported positive relations between the children’s book titles parents recognized and bilingual children’s emergent literacy. The most commonly used metric of print exposure is the Children’s Title Checklist (Sénéchal et al., 1998), which is used to determine the number of book titles a parent recognizes. In a study of Indian children with diverse L1s (e.g., Urdu, Hindi) in which 45% of children’s parents reported speaking both an Indian language and English at home, Kalia (2007) reported that exposure to English shared book reading was positively associated with bilingual children’s English oral language. Specifically, they found that the number of English book titles a parent recognized was associated with the complexity of the syntax produced by bilingual children, children’s knowledge regarding print concepts, narrative expression, and measures of phonological processing.

In a similar sample of Indian bilingual children receiving English instruction in kindergarten, Kalia and Reese (2009) investigated the impact of parents’ book-reading practices on children’s receptive vocabulary, phonological processing, and print concepts. Book-reading practices included the Children’s Title Checklist for India (Kalia, 2008), the number of children’s books, and the frequency of book reading at home. The Children’s Title Checklist was English specific, while the number of children’s books and book reading frequency were not language specific. The results of this study revealed a significant influence of Indian parents’ book-reading practices on their children’s English knowledge of print concepts and receptive vocabulary, which provided strong evidence of the crucial role of shared book reading in English in promoting Indian children’s oral language and emergent literacy development. Further, results indicated that high-levels of parents’ book-reading practices compensated for overall lower-levels of spoken English in children’s homes. Together, these studies of Indian bilingual children demonstrated the significant relation between second language (L2) parent-child shared book reading and L2 literacy development, particularly with respect to phonological processing, print concepts, and receptive vocabulary.

Intervention studies of parent-child shared reading have also provided evidence of early literacy enrichment (e.g., Chow et al., 2010; Tsybina & Eriks-Brophy, 2010). Chow et al., (2010) assigned bilingual Chinese-English kindergarteners to 3 different experimental groups (dialogic reading, typical reading, or the control group). They found that children in the dialogic reading group, who were scaffolded to speak more, use words actively, and engage in interactive conversations with their parents, showed enhanced phonological awareness in both their first language (L1) and L2. Furthermore, parent-child shared book reading promoted word reading skills in children’s L2 in both groups. Similarly, Tsybina and Eriks-Brophy (2010) found that children in the dialogic intervention group learned significantly more target L1 and L2 words than those in the control group. The target words were vocabulary items found in the books used in the study, including both nouns and verbs. Their sample consisted of bilingual Spanish heritage language preschoolers with expressive vocabulary delays in both L1 and L2. While no generalization was found (i.e., overall vocabulary gains did not significantly differ between groups), the results do suggest that parent-child shared book reading increased heritage language learners’ knowledge of those words embedded into the storybooks read. Finally, a meta-analysis of 54 studies of shared book reading interventions revealed an overall significant and positive effect of shared reading on English learners’ language and literacy skills (Fitton et al., 2018). However, it is notable that this meta-analysis included shared reading interventions that occurred beyond parent and child shared book reading, such as peer reading, service provider, researchers, and electronic book reading. Further, the meta-analysis made no distinction between bilingual subtypes, such as heritage language learners.

Socioeconomic Status Influences Parent-Child Shared Reading and Literacy

Culminating research reports that children from socioeconomically disadvantaged backgrounds have lower literacy abilities (Chatterji, 2006; Snow et al., 1998), fewer books (Bradley et al., 2001) and book choices (Neuman & Celano, 2001), and engage in parent-child shared book reading with less frequency (De Temple & Snow, 1996). Yet, monolingual parent-child shared book reading was not found to moderate the impact of socioeconomic status (SES) on children’s literacy (Bus et al., 1995; Raz & Bryant, 1990). Instead, Scarborough and Dobrich (1994) found that monolingual childhood SES and parent-child shared book reading each predicted children’s literacy outcomes. However, to date, no investigation has explicitly focused on the effect of parent-child book reading on the relation between childhood SES and early development of literacy skills among bilinguals. Preliminary evidence indicates that parent-child shared book reading may play a role in the relation between childhood SES and early literacy skills of heritage language learners.

One of the most commonly used proximal indices of SES is parents’ educational attainment (see Farah, 2017 for review). Farver, Xu, Lonigan, and Eppe (2013) found that parents’ educational attainment was positively associated with Spanish heritage language learner’s English expressive language, print knowledge skills, English literacy resources (i.e., print and educational materials), and parents’ English literacy involvement and literacy habits but was negatively associated with parents’ Spanish literacy involvement and literacy habits. For Indian bilingual children, Kalia and Reese (2009) found no significant relations between maternal education and children’s receptive vocabulary, phonological processing, or print concepts. However, there was a significant association between maternal educational levels and book-reading practices. Together, findings from these 2 studies suggest that in higher SES households—those with more educated mothers, more parent child reading takes place and there is more familiarity with popular children’s book titles, which may, as prior evidence suggests, enhance children’s oral language and emergent literacy skills. Thus, these 2 studies provided preliminary evidence of the potential role of parent-child book reading on the association between SES and bilingual children’s early literacy development.

In contrast, Hammer, Miccio, and Wagstaff (2003) reported no significant relation between parent-child shared book reading and Spanish heritage language learners’ emergent literacy skills. Specifically, they found that parents’ self-reported frequency of parent-child book reading did not predict children’s knowledge of contextual meaning, early literacy conventions, nor did it predict their understanding of the alphabet, letters, letter-sound correspondences, or concepts of print. The authors suggested that the lack of relations in their study may be due to the low level of family engagement in literacy activities and far fewer opportunities to access Spanish books for low-SES families. The variation of the findings among participants of different SES backgrounds could suggest that SES of bilingual families may have a larger effect on children’s early literacy development. The availability of literacy materials in certain languages may be limited and may cost more to access. Thus, it seems crucial to examine the effect of parent-child book reading and SES on children’s literacy skills.

The Current Study

The current study aimed to examine the potential benefits of parent-child shared book reading by first investigating its relation with the emergent literacy of heritage language learners from diverse language backgrounds. We hypothesized that parent-child shared book reading would be associated with children’s emergent literacy as found in numerous monolingual parent-child shared book reading studies. Second, we investigated the impact of parent-child shared book reading on the relation between SES and children’s emergent literacy. We hypothesized that parent-child shared book reading would influence the relation between SES and heritage language learners’ emergent literacy. Further, based on preliminary studies of parent-child shared book reading in specific SES strata, we hypothesized that parent-child shared book reading would mediate, rather than moderate the effect of SES on emergent literacy of heritage language learners.

Material and Methods

Participants

The present study includes participants from the Early Childhood Program Participation (ECPP) survey from the National Household Education Surveys Program of 2016 (NHES; NCES, 2017; McPhee et al., 2018). Developed by the National Center for Education Statistics (NCES), the NHES surveys were designed to collect nationally representative data directly from households regarding issues such as early childcare and education, participation in adult education, and parents’ involvement in children’s education. NHES uses a two-phase design in which (1) randomly sampled households completed a screening questionnaire, which was followed by (2) one person from each eligible household completing a topical survey. The NHES ECPP survey focused on children from birth to age 6 who were not yet enrolled in kindergarten. A total of 5,844 parents reported on childcare arrangements and educational programs, family learning activities, emergent literacy skills, as well as child, parent, and household characteristics. In the current study, parents of children younger than 2 years were excluded since they were not required to report their child’s early literacy abilities. Of the remaining 3,963 children (ages 2–6 years), children whose parents spoke a language other than English at home were identified as heritage language learners, which resulted in a sample of 1,005 children. After excluding 40 outliers in the sample, the final manuscript sample included 965 children between the ages of 2 and 6.

Measures

This study focuses on the impact of parent-child shared book reading on heritage language learners’ emergent literacy. We considered 2 factors to indicate the frequency of parent-child book reading, which included time the parent spent reading to the child and minutes during each of those times the parent spent reading to the child in the past week. We considered 7 variables to indicate children’s emergent literacy. This included children’s letter recognition, ability to write their name, if children read books on their own, if children read words in books, if children pretend to read words in books, as well as color identification. We considered 7 indicators as our proximal index of SES, which included parents’ highest education level, parents’ highest work status, community poverty percentage, total income, internet access, receipt of welfare benefits status, and house ownership. We also considered child characteristics, such as age and sex as covariates. Age and sex were selected a priori as covariates since previous literature has demonstrated that age (e.g., Vestheim et al., 2019) and sex (e.g., Logan & Johnston, 2010) impact the development of children’s literacy.

Emergent Literacy

Children’s emergent literacy was determined based on 5 questions posed to parents. To assess children’s letter recognition skills, parents were asked, “Can this child recognize the letters of the alphabet?” Parent’s response choices for the question were “No”, “Yes, Some of them”, “Yes, most of them”, and “Yes, all of them.” The second question was, “Can this child write his/her first name, even if some of the letters are backwards?” Parents could indicate either “No” or “Yes.” Third, parents were asked, “Does this child ever read or pretend to read storybooks on his/her own?” Parents could choose either “No” or “Yes.” The fourth question was, “Does this child actually read the words written in the book or does he/she look at the book and pretend to read?” Parents’ response choices were “Pretends to read”, “Actually reads the written words”, and “Does both.” The last question was recoded into 3 dummy variables indicating if the child pretended to read, actually read the written words, or did both. Lastly, children’s color identification skills were derived from the question, “Can this child identify the colors red, yellow, blue, and green by name?” Parents could choose “No”, “Yes, Some of them”, or “Yes, all of them.”

Parent-Child Shared Book Reading

The measure of the frequency of shared book reading was derived from 2 questions in the survey. Specifically, parents were asked, “How many times have you or someone in your family read to this child in the past week?” Parents’ response choices included either specifying the number of times or responding, “Not at all.” Parents were also asked, “About how many minutes on each of those times did you or someone in your family read to this child?” Similarly, parents then provided the number of minutes read during each book reading session described in the previous question. Notably, the questions were general to the frequency of shared book reading without specifying languages used during each book reading session. However, 37.6% of the parents reported speaking only a non-English language at home.

Socioeconomic Status (SES)

Parents’ Highest Educational Level.

Parents’ highest level of education was derived from 2 identical questions posed to both parents and we chose the highest degree obtained between both parents. Specifically, parents’ educational level was derived from the question: “What is the highest grade or level of school that this parent or guardian completed?” Parents were asked to choose from eleven possible response choices: “8th grade or less”, “High school but no diploma”, “High school diploma or equivalent (GED)”, “Vocational diploma after high school”, “Some college, but no degree”, “Associate’s degree (AA, AS)”, “Bachelor’s degree (BA, BS)”, “Some graduate or professional education, but no degree”, “Master’s degree (MA, MS)”, “Doctorate degree (PhD, EdD)”, and “Professional degree beyond bachelor’s degree (MD, DDS, JD, LLB).” We recoded this variable into 5 categories: less than high school, high school graduate or equivalent, vocational/technical education after high school, college graduate, and graduate or professional school.

Parents’ Highest Work Status.

Similarly, parents’ work status was derived from several questions. The first question was, “Which of the following best describes this person’s employment status?” Parents then marked one answer from “Employed for pay or income”, “Self-employed”, “Unemployed or out of work”, “Full-time student”, “Stay at home parent”, “Retired”, and “Disabled or unable to work.” If the parent chose “Employed for pay or income” or “Self-employed” in the first question, the parent was then asked, “(If employed or self-employed) about how many hours per week does he or she usually work for pay or income, counting all jobs?” Parents provided a specific number to this question. This response was used as an indicator of the parent’s work status either as “Working 35 hours or more per week” or “Working less than 35 hours per week.” If the parent chose “Unemployed or out of work” in the first question, the parent was instead asked, “(If unemployed or out of work) has this parent or guardian been actively looking for work in the past 4 weeks?” Response choices were “No” or “Yes.” If the parent indicated “Yes,” then his or her work status was coded as “Looking for work,” otherwise, the work status was coded as “Not in the labor force.” If the parent chose “Full-time student”, “Stay at home parent”, “Retired”, or “Disabled or unable to work,” then his or her work status was coded as “Not in the labor force.” We then chose the highest work status between the 2 parents to use as an indicator of parents’ highest work status.

Community Poverty Percentage.

This variable categorized the percentage of families in the sampled person’s community that have children under age 18 and had incomes in the 2011–2015 below the poverty line. There are 4 categories: less than 5%, 5% to 9%, 10% to 19%, and 20% or more.

Household Total Income.

Parents were asked, “Which category best fits the total income of all persons in your household over the past 12 months?” Parents’ response choices included: “$0 to $10,000”, “$10,001 to $20,000”, “$20,001 to $30,000”, “$30,001 to $40,000”, “$40,001 to $50,000”, “$50,001 to $60,000”, “$60,001 to $75,000”, “$75,001 to $100,000”, “$100,001 to $150,000”, and “$150,001 or more.”

Internet Access.

Parents were asked to choose either “No” or “Yes” to the question, “Do you have Internet access at home on a computer or tablet?”

Receipt of Welfare Benefits Status.

Parents were asked, “In the past 12 months, did your family ever receive benefits from any of the following programs?” Parents then marked “No” or “Yes” to each of the benefits, including “Temporary Assistance for Needy Families, or TANF”, “Your state welfare or family assistance program”, “Women, Infants, and Children, or WIC”, “Food Stamps”, “Medicaid”, “Child Health Insurance Program (CHIP)”, and “Section 8 housing assistance.” The responses were then combined to create a dummy variable. Parents who answered “Yes” to any of the questions were coded as 1 for this variable, and those who answered “No” were coded as 0.

House Ownership.

To indicate homeownership, parents were asked to choose one answer from “Owned or being bought by someone in this household”, “Rented by someone in this household, and “Occupied by some other arrangement.”

Children’s Demographics

Children’s demographic information included age, sex, race, and ethnicity. Parents were asked to report their child’s month and year of birth. When asked for their child’s sex, parents could choose either “Male” or “Female.” Parents were asked “What is this child’s race?” Parental response choices included “American Indian or Alaska Native”, “Asian”, “Black or African American”, “Native Hawaiian or other Pacific Islander”, and “White.” There was only one question regarding ethnicity. Parents were asked “Is this child of Hispanic, Latino, or Spanish origin?” Parents’ response choices were either “Yes” or “No.” In addition, we reported children’s parents’ responses to the question, “What language does this child speak most at home?” Parents then chose one response from “Child has not started to speak”, “English”, “Spanish”, “A language other than English or Spanish”, “English and Spanish equally”, and “English and another language equally.” We then recoded the responses of Spanish to be the same as “Another language and A language other than English.” The combined 4 categories of the responses were: child has not started to speak, English, a language other than English, and English and another language equally. We also reported children’s birthplace and their age when moved to the United States. Specifically, parents were asked, “Where was this child born?” Response choices included: “One of the 50 United States or the District of Columbia”, “One of the U.S. territories (Puerto Rico, Guam, American Samoa, U.S. Virgin Islands, or Mariana Islands)”, and “Another country.” For parents choosing either of the latter 2 responses, they were then asked, “How old was this child when he/she first moved to the 50 United States or the District of Columbia?” A numerical age was then provided by the parents.

Heritage Language Learner Status

Since heritage language learners are defined as children who grow up in families where a non-English language is spoken, children’s heritage language learner status was determined by the language their parents reported speaking at home. Specifically, parents were asked, “Which language(s) are spoken at home by the adults in this household?” Parents then marked all languages that applied, including “English”, “Spanish or Spanish Creole”, “French (including Patois, Creole, Cajun)”, “Chinese”, and “Other languages: Specify.” Children whose parents responded with at least one language other than English were included.

Analytic Overview

A Confirmatory Factor Analysis (CFA) was used to examine the factor structure of parent-child book reading, SES, and emergent literacy. Parent-child book reading included the times spent reading to the child and minutes spent reading together during each reading session. SES consisted of parents’ highest-grade level completed, parents’ highest work status, community poverty percentage, household total income, internet access, receipt of welfare benefits status, and house ownership. Emergent literacy included letter recognition, name writing, reading books on his or her own, and reading words in books, pretending to read words in books, as well as color identification.

First, to answer our primary research question, we used latent variable path analyses (LVPA) to estimate the effect of parent-child shared book reading on heritage language learners’ emergent literacy. LVPA allowed us to combine measures of emergent literacy on different scales. In the first model, the latent emergent literacy variable was predicted by the latent parent-child shared book reading variable. Age and sex were used as control variables in the model. To make population inferences, weighting is used in the NHES study to account for differential probabilities of selection and to reduce potential bias due to nonresponse and differential coverage of subpopulations. Therefore, a survey weighting variable was included in the model to take into account the complex survey design of the NHES data.

Second, we aimed to investigate if parent-child shared book reading served as either a moderator or mediator of the relation between SES on heritage language learners’ emergent literacy. To determine whether parent-child shared book reading mediated the relation between childhood SES and children’s early literacy, we tested the effects of SES on emergent literacy and parent-child book reading using LVPA. We then built on our previous LVPA model, adding the SES latent variable as an independent variable and including parent-child shared book reading as a mediator. Finally, to investigate the moderating role of parent-child book reading on the association between SES and children’s emergent literacy, we constructed a model with SES, parent-child book reading, as well as their interaction term predicting emergent literacy.

Follow-up analyses were performed to evaluate whether the path from parent-child book reading to children’s emergent literacy differed between younger and older children, as well as whether the role of SES differed in either age group. To conduct follow-up analyses, we divided our sample into 2 groups, younger children (ages 2–4 years) and older children (ages 4–6 years). All LVPA models were conducted in Mplus 8 (Muthén & Muthén, 2017).

Method of Estimation, Sample Size, Missing Data & Outliers

We used weighted least squares with means and variances adjusted (WLSMV) as an estimator in the analyses. The WLSMV is a robust estimator, which can account for non-normal variables and can best deal with categorical variables (Brown, 2006). Multivariate outliers were detected using Cook’s distance (Cook, 1977), which represents a statistic that combines standardized residuals and leverage values to flag influential observations. We obtained each observation’s Cook’s distance in Mplus and eliminated observations with Cook’s distance more than 1. A total of 40 outliers were found and removed from the analysis, leaving a final sample of 965 individuals in the LVPA models.

Goodness-of-Fit Test

The following model fit goodness of fit tests were reported: root mean square error of approximation (RMSEA; Browne & Cudeck, 1993) with 90% Confidence Intervals (CIs) and comparative fit index (CFI; Bentler, 1990). To indicate a good model fit we evaluated them against the following guidelines: the RMSEA estimate should be lower than 0.05 (Hu & Bentler, 1999); CFI should be higher than 0.90 (Bentler & Bonett, 1980).

Results

Descriptive Statistics

Demographic information and descriptive statistics for the full sample are presented in Table 1. Demographic information by SES can also be found in Table 1. Half of the children were girls (48.2%), and more than half of the full sample were White (59.2%). Among all children, 876 (90.8%) were born in the United States, 12 (1.2%) were born in one of the U.S. territories, and 77 (8%) were born in another country. For those who were born in a U.S. territory or another country, the age of the child when they first moved to the U.S. ranged from 0 to 4 years old. Of the 77 children who were born outside of the U.S. and U.S. territories, 9 moved to the U.S. before they were one year old, 25 moved at one year old, 23 moved at 2 years old, 25 moved at 3 years old, and 7 moved at 4 years old. Table 2 provides the languages spoken at home by parents and children. About 37.6% of parents spoke only a non-English language at home, while 62.4% of parents reported speaking a non-English language and English at home. A total of 25.9% of children were reported to speak only a non-English language at home, 30.2% of children were reported to speak a non-English language and English at home equally, and 40.9% of the children were reported to speak mostly English at home. Consistent with previous research on heritage language learners, parents of heritage language learners are more likely to continue to use their native language at home while the heritage language learners demonstrate mixed use of their heritage and the majority ethnolinguistic languages (Jee, 2018; Liang, 2018; Zhang & Koda, 2011).

Table 1.

Demographic Information and Descriptive Statistics (unweighted)

Variable	Full Sample (N=965)	Low-SESa (N=285)	Mid-SESb (N=307)	High-SESc (N=373)
Age in months, mean (SD), range 24–79	45.50 (12.40)	45.45 (12.59)	45.27 (12.23)	45.73 (12.42)
Female (%)	465 (48.2)	143 (50.2)	145 (47.2)	177 (47.5)
American Indian or Alaska Native (%)	36 (3.7)	13 (4.6)	12 (3.9)	11 (2.9)
Asian (%)	238 (24.7)	32 (11.2)	59 (19.2)	147 (39.4)
Black or African American (%)	69 (7.2)	26 (9.1)	23 (7.5)	20 (5.4)
Native Hawaiian or other Pacific Islander (%)	10 (1.0)	4 (1.4)	2 (0.7)	4 (1.1)
White (%)	571 (59.2)	160 (56.1)	192 (62.5)	219 (58.7)
Spanish, Hispanic, or Latino (%)	512 (53.1)	210 (73.7)	182 (59.3)	120 (32.2)
Hispanic-race not reported (%)	124 (12.8)	72 (25.3)	38 (12.4)	14 (3.8)
Recognize letters of alphabet (%)
No	162 (16.8)	74 (26.0)	53 (17.3)	35 (9.4)
Some of them	275 (28.5)	106 (37.2)	88 (28.7)	81 (21.7)
Most of them	187 (19.4)	46 (16.1)	70 (22.8)	71 (19.0)
All of them	341 (35.3)	59 (20.7)	96 (31.3)	186 (49.9)
Ability to write first name (%)	415 (43.0)	92 (32.3)	134 (43.6)	189 (50.7)
Read books by him/herself (%)	817 (84.7)	225 (78.9)	254 (82.7)	338 (90.6)
Actually read the written words in the book (%)	44 (4.6)	9 (3.2)	12 (3.9)	23 (6.2)
Pretend/actually read the written words in the book (%)	128 (13.3)	21 (7.4)	39 (12.7)	68 (18.2)
Identify colors by name (%)
No	58 (6.0)	32 (11.2)	18 (5.9)	8 (2.1)
Some of them	187 (19.4)	79 (27.7)	54 (17.6)	54 (14.5)
All of them	720 (74.6)	174 (61.1)	235 (76.5)	311 (83.4)
Time spent reading to child, mean (SD), range 0–30	5.88 (5.06)	4.30 (4.18)	5.32 (4.29)	7.53 (5.74)
Minutes spent each time reading to child, mean (SD), range 0–75	16.08 (11.06)	14.97 (11.56)	16.69 (12.15)	16.42 (9.59)
Books child own, mean (SD), range 0–800	51.32 (72.8)	29.87 (52.22)	42.30 (57.07)	75.14 (89.37)
Visited a library in the past month (%)	374 (38.8)	87 (30.5)	104 (33.9)	183 (49.1)
Visited a bookstore in the past month (%)	354 (36.7)	94 (33.0)	100 (32.6)	160 (42.9)
Parent highest education level (%)
Less than high school	81 (8.4)	59 (20.7)	21 (6.8)	1 (0.3)
High school graduate or equivalent	117 (12.1)	68 (23.9)	43 (14.0)	6 (1.6)
Vocational/technical education after high school or some college	241 (25.0)	92 (32.3)	107 (34.9)	42 (11.3)
College graduate	220 (22.8)	36 (12.6)	85 (27.7)	99 (26.5)
Graduate or professional school	306 (31.7)	30 (10.5)	51 (16.6)	225 (60.3)
Parent highest work status (%)
Not in labor force	54 (5.6)	36 (12.6)	13 (4.2)	5 (1.3)
Looking for job	26 (2.7)	21 (7.4)	1 (0.3)	4 (1.1)
Less than 35 hours per week	82 (8.5)	54 (18.9)	14 (4.6)	14 (3.8)
35 hours or more per week	803 (83.2)	174 (61.1)	279 (90.9)	350 (93.8)
Community poverty percentage (%)
20% or more	121 (12.5)	64 (22.5)	39 (12.7)	18 (4.8)
10% to 19%	194 (30.5)	133 (46.7)	107 (34.9)	54 (14.5)
5% to 9%	267 (27.7)	66 (23.2)	97 (31.6)	104 (27.9)
less than 5%	283 (29.3)	22 (7.7)	64 (20.8)	197 (52.8)
Internet access (%)	851 (88.2)	206 (72.3)	278 (90.6)	367 (98.4)
Receipt of welfare benefits status (%)	435 (45.1)	253 (88.8)	161 (52.4)	21 (5.6)
House ownership (%)
Owned or being bought by someone in this household	555 (57.5)	84 (29.5)	180 (58.6)	291 (78.0)
Rented by someone in this household	380 (39.4)	185 (64.9)	118 (38.4)	77 (20.6)
Occupied by some other arrangement	30 (3.1)	16 (5.6)	9 (2.9)	5 (1.3)

Note. N = 965.

^a= total family income between $0 and $30,000.

^b= total family income between $30,001 and $75,000.

^c= total family income above $75,001.

Table 2.

Languages Spoken at Home

Variable	N	Percent
Adults
Speak non-English language(s) only	363	37.6
Spanish only	222	23.0
French only	2	0.2
Chinese only	24	2.5
Other only	113	11.7
2 non-English languages	2	0.2
Speak English and 1 non-English language	568	58.9
Speak English and 2 non-English languages	32	3.3
Speak English, Spanish, French, & Other	1	0.1
Speak English, Spanish, Chinese, & Other	1	0.1
Children
Has not started to speak	29	3
English	395	40.9
A language other than English	250	25.9
English and another language equally	291	30.2

Note. N = 965.

Preliminary Analysis

Table 3 provides correlations for all of the observed independent and dependent variables included in the models containing latent variables. The variables included in the parent-child shared book reading latent factor were significantly correlated (r = 0.254, p < .001). All the variables included in the emergent literacy latent construct were significantly correlated (p < .01). Likewise, all SES indicators were significantly correlated (p < .001), except for internet access, which was not correlated with parent highest work status. For all the SES variables, we checked the variance inflation factors (VIFs) to identify any collinearity issues. Using the criteria that a value of 10 indicates a collinearity problem (O’Brien, 2007) and a value of 5 indicates considerable collinearity (Menard, 2001), we found no evidence of a collinearity issue. The SES variables VIFs ranged from 1.09 to 2.75.

Table 3.

Correlation Matrix of Observed Variables

	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15
1. Time	1
2. Minutes	.254	1
3. Education	.294	.099	1
4. Work	.063	−.041	.159	1
5. Community	.189	.06	.421	.156	1
6. Income	.288	.047	.609	.337	.476	1
7. Internet	.128	.022	.346	.046	.257	.343	1
8. Benefit	−.263	−.084	−.518	−.275	−.404	−.705	−.268	1
9. House	.111	.027	.181	.134	.208	.394	.115	.354	1
10. Letter	.204	.184	.283	.107	.173	.291	.149	.282	.119	1
11. Name	.066	.143	.092	.068	.107	.147	.059	.122	.038	.553	1
12. Books	.190	.171	.099	.106	.103	.142	−.004	.158	.065	.185	.166	1
13. Words	.053	.083	.073	.016	.012	.064	−.012	.058	.063	.231	.232	.093	1
14. Pretend	.166	.091	.136	.017	0.08	.120	.058	.072	0.024	.289	.265	.166	−.085	1
15. Color	.175	.191	.181	.092	.184	.232	.178	.242	.134	.538	.395	.246	.118	.196	1

Note. Correlations greater than 0.107 are significant at p < .001 level. Correlations greater than 0.083 are significant at p < .01 level. Correlations greater than 0.063 are significant at p < .05 level. Time = Time spent reading to child; Minutes = Minutes spent each time reading to child; Education = Parent highest education level; Work = Parent highest work status; Community = community poverty percentage; Income = Total income; Internet = Internet access; Benefit = Receipt of welfare benefits status; House = House ownership; Letter = Recognize letters of alphabet; Name = Ability to write first name; Books = Read books by him/herself; Words = Actually read the written words in the book; Pretend = Pretend/Actually read the written words in the book; Color = Identify colors by name.

Confirmatory Factor Analysis (CFA)

A CFA model was constructed first to test the adequacy of the measurement model. Since the factor loading of one variable (pretend to read words in books) was low (β = 0.052), this variable was excluded from the final CFA model. The resulting model showed an excellent fit to the data, χ² (87) = 294.723, p < .001; CFI = 0.960; RMSEA = 0.050, 90% CI [0.044, 0.056]. All observed variables had fair to excellent factor loadings to the latent variables (Figure 1). Therefore, the CFA model demonstrated that the remaining observed variables were adequate indicators of the latent constructs of (1) parent-child book reading, (2) socioeconomic status, and (3) emergent literacy.

Figure 1.

Confirmatory Factor Analysis of Parent-Child Book Reading, Socioeconomic Status, and Emergent Literacy

Note. Time = Time spent reading to child; Minutes = Minutes spent each time reading to child; Education = Parent highest education level; Work = Parent highest work status; Community = community poverty percentage; Income = Total income; Internet = Internet access; Benefit = Receipt of welfare benefits status; House = House ownership; Letter = Recognize letters of alphabet; Name = Ability to write first name; Books = Read books by him/herself; Words = Actually read the written words in the book; Pretend = Pretend/Actually read the written words in the book; Color = Identify colors by name. All coefficients are standardized. All paths significant at p < .001.

Relation Between Parent-Child Shared Book Reading and Heritage Language Learners’ Emergent Literacy

The first LVPA model included parent-child book reading as a latent construct predicting heritage language learners’ emergent literacy. The child’s sex and age were included as covariates. The model provided a good fit to the data, χ² (32) = 196.758, p < .001; CFI = 0.910; RMSEA = 0.073, 90% CI [0.063, 0.083]. Figure 2 shows the significant standardized model coefficients. The analysis revealed that the path from parent-child book reading to children’s emergent literacy was positive and significant (β = 0.455, p < .001). This indicates that parent-child shared book reading significantly contributed to heritage language learners’ emergent literacy. Together, the covariates and latent variables accounted for 68.7% of the variance in emergent literacy.

Figure 2.

LVPA Model of Parent-Child Book Reading Predicting Children’s Emergent Literacy

Note. LVPA model of parent-child book reading predicting emergent literacy. Emergent literacy is a latent variable consisted of 6 observed variables: Letter, Name, Books, Words, Pretend, Color. Letter = Recognize letters of alphabet; Name = Ability to write first name; Books = Read books by him/herself; Words = Actually read the written words in the book; Pretend = Pretend/Actually read the written words in the book; Color = Identify colors by name. Parent-Child Book Reading is a latent variable composed of Times and Minutes. Time = Time spent reading to child; Minutes = Minutes spent each time reading to child. All coefficients are standardized. *p < .05; ***p < .001.

Follow-up LVPA models were run after splitting the data into a younger (2–4 years old) and an older (4–6 years old) sample. Both LVPA models provided good fit to the data (younger sample: χ² (31) = 137.782, p < .001; CFI = 0.909; RMSEA = 0.080, 90% CI [0.066, 0.094]; older sample: χ² (32) = 87.309, p < .001; CFI = 0.928; RMSEA = 0.064, 90% CI [0.048, 0.080]). Both models showed that the path from parent-child book reading to children’s emergent literacy was positive and significant (younger sample: β = 0.658, p < .001; older sample: β = 0.436, p < .001). Together, the covariates and latent variables in the younger sample accounted for 72.7% of the variance in emergent literacy, while the covariates and latent variables in the older sample accounted for 46.2% of the variance in emergent literacy. This suggests that parent-child book reading is significantly related to emergent literacy for children in both age groups, although we note that the effect of parent-child book reading was stronger for younger children.

The Influence of Parent-Child Shared Book Reading on the Relation between Socioeconomic Status and Heritage Language Learners’ Emergent Literacy

Prior to implementing the LVPA mediation model, one LVPA model was conducted with SES predicting children’s emergent literacy and parent-child shared book reading. The model had a good fit to the data, χ² (87) = 294.723, p < .001; CFI = 0.960; RMSEA = 0.050, 90% CI [0.044, 0.056]. SES significantly predicted children’s emergent literacy (β = 0.352; p < .001) and the frequency of parent-child book reading (β = 0.542; p < .001). To answer our main question of interest, a LVPA mediation model was constructed and provided a good fit to the data, χ² (115) = 354.001, p < .001; CFI = 0.957; RMSEA = 0.046, 90% CI [0.041, 0.052]. Figure 3 displays the significant standardized model coefficients. SES was significantly directly (path c: β = 0.392; p < .001) and indirectly (path a-b: β = 0.121, p < .001, 95% CI [0.064, 0.177]) related to children’s emergent literacy. The indirect relation was mediated by parent-child shared book reading. Moreover, the direct relation between SES and emergent literacy remained significant (path c’: β = 0.272; p < .001), even when parent-child shared book reading acted as a mediator. Thus, parent-child book reading partially mediated the relation between SES and children’s emergent literacy. Together, the control and latent variables accounted for 65.2% of the variance in emergent literacy.

Figure 3.

LVPA Model of Parent-Child Shared Book Reading Mediating the Relation between SES and Children’s Emergent Literacy

Note. LVPA mediation model of parent-child book reading mediating the relation between SES and children’s emergent literacy. Parent-Child Book Reading is a latent variable composed of Times and Minutes. Time = Time spent reading to child; Minutes = Minutes spent each time reading to child. SES is a latent variable composed of the variables: Education, Work, Community, Income, Internet, Non-benefit, and House. Education = Parent highest education level; Work = Parent highest work status; Community = community poverty percentage; Income = Total income; Internet = Internet access; Non-benefit = Non-Receipt of welfare benefits status; House = House ownership. Emergent literacy is a latent variable consisted of 6 observed variables: Letter, Name, Books, Words, Pretend, Color. Letter = Recognize letters of alphabet; Name = Ability to write first name; Books = Read books by him/herself; Words = Actually read the written words in the book; Pretend = Pretend/Actually read the written words in the book; Color = Identify colors by name. All coefficients are standardized. *p < .05; ***p < .001.

Follow-up analysis with the younger and older sample showed similar results to the whole group analysis. LVPA models provide good fit to the data (younger sample: χ² (115) = 284.755, p < .001; CFI = 0.946; RMSEA = 0.052, 90% CI [0.045, 0.060]; older sample: χ² (115) = 203.820, p < .001; CFI = 0.959; RMSEA = 0.043, 90% CI [0.033, 0.052]). In both the younger and older samples, SES was significantly related to children’s emergent literacy directly (younger sample: β = 0.453; p < .001; older sample: β = 0.427; p < .001) and indirectly (younger sample: β = 0.203; p = .001, 95% CI [0.081, 0.325]; older sample: β = 0.104; p = .003, 95% CI [0.035, 0.174]). In both samples, the indirect relation was mediated by parent-child shared book reading. Moreover, the direct relation between SES and emergent literacy remained significant (younger sample: β = 0.250; p = .001; older sample: β = 0.323; p < .001), even when parent-child shared book reading acted as a mediator. Together, the covariates and latent variables in the younger sample accounted for 57.1% of the variance in emergent literacy, while the covariates and latent variables in the older sample accounted for 48.7% of the variance in emergent literacy. Thus, results suggest that parent-child shared book reading partially mediated the relation between SES and children’s emergent literacy for both the younger and older samples, although we note that the mediating effect of parent-child book reading was stronger for younger children.

An additional LVPA model was specified to investigate the moderating effect of parent-child shared book reading on the relation between SES and children’s emergent literacy and had good model fit to the data, χ² (129) = 580.512, p < .001; CFI = 0.918; RMSEA = 0.060, 90% CI [0.055, 0.065]. However, the interaction of SES and parent-child shared book reading was not significant (p = .489). Therefore, no evidence was found to support a moderating effect of parent-child shared book reading on the relation between SES and children’s emergent literacy. As the results were not significant, no follow-up analyses were conducted.

Discussion

Parent-child shared book reading is a critical component of an enriched home literacy environment. It provides the very foundation by which children acquire literacy. Research has provided substantial evidence confirming the importance of parent-child shared book reading in monolingual children’s literacy development (e.g., Bus et al., 1995; Mol & Bus, 2011; Scarborough & Dobrich, 1994). However, given the 12 million children speaking a variety of languages other than English in the United States, the relative sparsity of literature on the relationship between parent-child shared book reading and linguistically diverse children’s literacy development is quite astounding. Heritage language learners constitute the majority of linguistically diverse children in United States schools; yet, little research attention has focused on this bilingual subpopulation. In the current study, we investigated the relation between parent-child shared book reading and emergent literacy in a large, national sample of heritage language learners at the very early stages of acquiring literacy, from ages 2 to 6 years old.

While previous studies mostly focused on heritage language learners from a specific heritage language background, such as Spanish (e.g., Bitetti & Hammer, 2016), the current study is the first to investigate this relation in a nationally representative sample of heritage language learners from diverse home language backgrounds (i.e., Spanish, French, Chinese, and other languages). There are 2 key findings that emerged from the current study. First, parent-child shared reading was related to heritage language learners’ emergent literacy. Second, parent-child shared book reading mediated, but did not moderate, the relation between SES and heritage language learners’ emergent literacy. Below we discuss these findings in the context of the current literature.

Heritage Language Learners’ Emergent Literacy

Emergent literacy in monolinguals and bilinguals has been linked to parent-child shared book reading. In the current study, we found that parent child shared book reading was positively associated with heritage language learners’ emergent literacy. By investigating a national sample of heritage language learners, we were able to uncover that not only is there a relation between parent-child shared book reading and children’s emergent literacy but also that this relation is stronger in younger heritage language learners (ages 2–4 years) compared to older heritage language learners (ages 4–6 years). This finding is consistent with monolingual parent-child book reading literature indicating stronger relation between parent-child book reading and emergent literacy for younger samples (Bus et al., 1995). It is possible that older children are more likely to be encouraged to read independently, and thus, their need for parent-child shared book reading may be attenuated by the increase in their developing emergent literacy skills. While measurement of the frequency of parent-child shared book reading and children’s emergent literacy skills prevent determination of a causal relationship, we can eliminate several alternative explanations of our results. Notably, none of the children in this study were, at the time of the study, attending school, thus, allowing us to rule out the acquisition of emergent English literacy at school. We were also able to rule out previously reported sex and age differences by including age and sex as covariates in our analyses.

Parent-Child Shared Book Reading Mediates the SES and Emergent Literacy Relation

A plethora of research has linked SES to children’s early literacy development (e.g., Carroll et al., 2019; Van Steensel, 2006). Monolingual socioeconomically disadvantaged children are thought to benefit as much, if not more, from parent-child shared book reading than children growing up in middle and high SES families (e.g., Bus et al., 1995; Raz & Bryant, 1990). Prior to the current study, the relation between SES and heritage language learners’ parent-child shared book reading had not yet been the focus of an investigation. Here, we report that the relation between SES and heritage language learners’ emergent literacy was partially mediated by parent-child shared book reading. In other words, parent-child shared book reading can explain, in part, the relation between SES and heritage language learners’ early literacy ability. This suggests that families with higher SES are more likely to engage in parent-child shared book reading, which is associated with better literacy development among heritage language learners.

The mediating effect of parent-child shared book reading on the relation between SES and heritage language learners’ emergent literacy was stronger for younger children than older children. Moreover, similar to the results of the monolingual child literature (Bus et al., 1995; Raz & Bryant, 1990), there was no moderating effect of parent-child book reading on the relation between SES and heritage language learners’ emergent literacy. Our results suggest that while parent-child book reading can act, in part, as an intermediary between SES and heritage language learners’ early literacy, parent-child book reading does not alter the strength of the relation between SES and heritage language learners’ early literacy.

Limitations and Implications

Our findings have implications for informal educational “at-home” practices. First, for young children in families speaking a heritage language, parents are encouraged to read frequently to their children. The frequency with which children are read to is related to their emergent literacy skills. Second, although heritage language learners from higher socioeconomic backgrounds tended to be read to more frequently, the benefit of parent-child shared book reading was evident regardless of socioeconomic background. Therefore, literacy programs promoting shared book reading practices at home are needed, especially for children from low SES backgrounds.

While the results from studies of monolingual parent-child book reading are positive and highly replicable, the extent of the effect of parent-child book reading appears to be dependent upon measurement. Studies using a single question asking parents about the frequency of book reading revealed weaker correlations with oral language and basic reading skills than studies using print exposure checklists (Mol & Bus, 2011). In the current study, information about the parent-child book reading, socioeconomic status, and children’s emergent literacy was the result of parental report rather than direct observation or assessment. Therefore, caution is warranted in interpreting the findings as the reported values may be diminished as has been shown in monolingual studies. Alternatively, values may be inflated as the results may be affected by social desirability bias, which is a general tendency to report socially desirable behaviors (Chung & Monroe, 2003; Presser & Stinson, 1998). Future studies may wish to consider using shared book reading parent reported measures as well as home observations. Additionally, the most commonly used metric of print exposure is the Children’s Title Checklist (Sénéchal et al., 1998); however, the checklists are based on the popularity of children’s books, which are likely to be biased towards a monolingual, white, and more affluent population. Thus, future studies may wish to take the approach used by Kalia (2007) and modify the children’s book title checklist measure for use as an indicator of the heritage language learners’ book reading practices. In terms of outcome measures, future research may wish to include direct measures of children’s emergent reading skills.

In addition, the languages used by parents during shared book reading practices in the current study are unknown, although the information about languages spoken by parents at home is available (see Table 2). Future studies may want to consider collecting this information to gain a better understanding of language-specific links between bilingual parent-child book reading and children’s emergent literacy. Likewise, information about whether the books the child owned were in English or other languages is not available. A previous study reported no association between parent-child book reading and bilingual children’s literacy; the author attributed the outcome to limited access to Spanish books (Hammer et al., 2003). However, the number of bilingual books on the United States market appears to be increasing (Ahuile, 2019). As such, future studies may wish to explore the language-specific relations between book access in different languages and children’s literacy outcomes. Moreover, one cannot assume that parents read books in the language the text provided (e.g., books written in English were read by parents to their children in English). It is highly likely that Spanish-speaking parents read bilingual books aloud to their child using their most comfortable language – Spanish (Wessels, 2014). As such, to disentangle the complex relations between parent-child book reading and book texts, researchers may wish to use observational studies of parent-child book reading.

Summary and Conclusions

The importance of parent-child book reading for monolingual children has been stressed in a significant body of research and integrated into educational practices aimed at enriching home literacy environment. Yet, this relation between parent-child book reading and heritage language learners’ emergent literacy has received little research attention to date. The current study uses data from a national sample to fill an important gap in the literature on early literacy development among heritage language learners. First, we report that heritage language learners’ emergent literacy was associated not only with parent-child shared book reading but also with SES. This finding extends the current literature on emergent literacy and has important implications for both research and education practice. Second, we show, for the first time, that the relation between SES and heritage language learners’ early literacy skills are mediated by parent-child shared book reading. Taken together, these results provide new and critical insights into early literacy development among heritage language learners.

Highlights.

• Parents-child book reading fostered heritage language learners’ emergent literacy

• Socioeconomic status predicts heritage language learners’ emergent literacy

• Parent-child book reading partially explain the SES-emergent literacy relation