Influences of the Home Language and Literacy Environment on Spanish and English Vocabulary Growth among Dual Language Learners

Abstract

Little is known regarding the extent to which aspects of the home language and literacy environment (HLE) promote growth in language skills among dual language learners (DLLs). Therefore, the purpose of this study was to evaluate which aspects of the HLE significantly predict growth in English and Spanish vocabulary among Spanish-speaking DLLs. 944 Spanish-speaking DLLs (51.6% female; mean age = 53.77 months) completed assessments of English and Spanish vocabulary at four time points across two academic years. Parents completed a survey of the HLE that included information on language exposure, reading exposure, and family reading habits. Results indicated that specific literacy-related practices, including availability of books in the home, language read to the child, and parental reading skills were significant predictors of growth in children’s Spanish and English vocabulary knowledge, even after controlling children’s initial level of language skills and family socioeconomic status.

Children’s early language skills in preschool predict reading development in the elementary years (e.g., Murphy, Justice, O’Connell, Pentimonti, & Kaderavek, 2016; Scarborough, 2001; Whitehurst & Lonigan, 1998). Consequently, it is important to understand the factors that promote acquisition of language skills early in life. Although a substantial body of evidence indicates that the home language and literacy environment (HLE) plays an important role in the development of early language skills and school readiness among monolingual children (e.g., Forget-Dubois, Dionne, Lemelin, Pérusse, Tremblay, & Boivin, 2009; Payne, Whitehurst, & Angell, 1994; Son & Morrison, 2010), fewer studies have examined how the HLE influences bilingual language development for children exposed to more than one language (i.e., dual language learners [DLLs]; Hammer, Hoff, Uchikoshi, Gillanders, Castro, & Sandilos, 2014). DLLs often have lower levels of proficiency in both of their languages than do their monolingual peers (e.g., Goodrich & Lonigan, 2018; Hoff, Core, Place, Rumiche, Señor, & Parra, 2012; Vagh, Pan, & Mancilla-Martinez, 2009); however, these deficits may not be indicative of atypical language development but may be due to differences in language exposure between bilingual and monolingual children (Bedore & Peña, 2008). Nevertheless, despite an overall reduction in the achievement gap between monolingual children and DLLs over the course of the last decade (Kieffer & Thompson, 2018), DLLs have lower reading achievement throughout school than do their monolingual peers. Understanding factors that promote successful bilingual language acquisition in early childhood will assist in identifying at-risk children and inform interventions designed to prevent them from falling behind their peers in language and literacy development. The extant evidence on relations between aspects of the HLE and children’s language skills among monolingual children and DLLs comes from both concurrent (e.g., Dixon, Zhao, Quiroz, & Snow, 2012) and longitudinal (Forget-Dubois et al., 2009) designs; however, no studies of DLLs have examined how the HLE influences growth in children’s developing language skills. Consequently, the purpose of this study was to identify aspects of the HLE (i.e., language input and exposure in Spanish and English, exposure to print in Spanish and English, and family reading habits) that predict initial level and rate of growth of Spanish-speaking DLLs’ vocabulary knowledge in Spanish and English.

Theoretical Framework

According to Bronfenbrenner’s ecological systems theory (Bronfenbrenner, 2005), the specific microsystems (i.e., environments) in which children exist have a direct influence on development. Moreover, the ecological systems framework posits that the interactions between children and their environment (including other people in the environment [e.g., caregivers]) have a particularly important influence on development early in life. During the early childhood years, the home environment represents the predominant microsystem for many children. Thus, experiences children have at home should play a primary role in early language development, as exposure to conversational speech and structured language experiences, such as shared book reading, frequently occur at home. Indeed, research indicates that exposure to both spoken and written language at home during the preschool years influence language development. For example, language exposure plays a key role in language development (e.g., Hart & Risley, 1995; Huttenlocher, Haight, Bryk, Seltzer, & Lyons, 1991). Specifically, children exposed to greater amounts of adult speech have higher levels of language proficiency and faster rates of growth of various aspects of language skills, including vocabulary knowledge (Huttenlocher et al., 1991), syntactic development (Huttenlocher, Waterfall, Vasilyeva, Vevea, & Hedges, 2010), and narrative skills (Scheele, Leseman, Mayo, & Elbers, 2012). Similarly, home literacy practices influence children’s developing language skills. Factors including parental reading ability, exposure to books in the home, and shared reading experiences are positively associated with early language development (Payne et al., 1994).

Although research has consistently identified language exposure (Rowe, 2012), exposure to print (e.g., Caglar-Ryeng, Eklund, & Nergård-Nilssen, 2020), and family reading habits/beliefs (Weigel, Martin, & Bennett, 2006) as aspects of the HLE that predict language outcomes for monolingual children, less is known about how the HLE influences language development for DLLs. Ecological systems theory (Bronfenbrenner, 2005) also can be applied to describe how the HLE influences language development for DLLs. In environments in which children hear two languages on a regular basis, children’s exposure to language and reading-related activities is distributed across those languages, and the specific degree to which this distribution occurs varies across children and families. For example, Spanish-English DLLs growing up in a bilingual household may speak Spanish with one caregiver and English with another, with substantial experience in both languages occurring from birth. In this context, input and exposure to language and literacy activities may be relatively balanced across Spanish and English throughout early childhood. In contrast, some Spanish-English DLLs growing up in the U.S. may live with caregivers who recently immigrated to the U.S. from Latin America and predominantly speak Spanish at home. In this context, input and exposure to language and literacy activities may occur primarily in Spanish. Distribution of language input and exposure to literacy activities across languages yields greater heterogeneity in the nature of the HLE in dual-language contexts relative to monolingual contexts. Moreover, DLLs disproportionately come from lower socioeconomic backgrounds (Park, O’Toole, & Katsiaficas, 2017), and socioeconomic status is associated with language development (Hart & Risley, 1995; Pace, Luo, Hirsh-Pasek, & Golinkoff, 2017). Therefore, theoretical predictions of the influence of the HLE on DLLs’ language development must consider the complexity of DLLs’ HLE and ensure that associations between the HLE and language development for DLLs are independent of the influence of socioeconomic status.

Prior evidence indicates the influence of the HLE for Spanish-speaking DLLs is stronger within than across languages (Scheele, Leseman, & Mayo, 2010). Specifically, Scheele et al. reported that storytelling and conversations in L1 consistently influenced language outcomes in L1 but not L2. Other research has supported the finding that influences of the HLE are language-specific for young DLLs (e.g., Dixon et al., 2012; Farver et al., 2013; Patterson, 2002; Quiroz, Snow, & Zhao, 2010), as percent of language exposure in English is positively correlated with English language development and negatively correlated with Spanish language development (Parra, Hoff, & Core, 2011), and input from native speakers results in greater language development than does input from non-native speakers (Hoff, Giguere, Quinn, & Lauro, 2018). Additionally, evidence indicates that increased exposure to and use of Spanish in the home is a positive predictor of children’s Spanish oral language, whereas English oral language is primarily influenced by children’s use of English (Lewis, Sandilos, Hammer, Sawyer, & Méndez, 2016). In sum, theory and prior evidence converge to suggest that influences of the HLE on child language development are language-specific for DLLs.

Defining the HLE

According to Sénéchal and LeFevre’s (2002; 2014) home literacy model, children are exposed to both informal and formal literacy experiences at home. Informal experiences refer to interactions in which print is present but is not the focus of the interaction, whereas formal literacy activities involve specific focus on printed text. In the context of the home literacy model, shared storybook reading would be considered an informal literacy activity. Although shared storybook reading may initially appear to be specifically focused on print, evidence indicates that children often do not fixate on the printed text directly, instead attending primarily to pictures and illustrations in the text (e.g., Evans, Saint-Aubin, & Landry, 2009). Thus, the print itself is not the focus of the shared reading experience for the child, even though it is present during shared reading. Formal literacy activities in the home involve instances in which parents directly teach literacy skills such as letter knowledge, reading, or writing. Predictions derived from the home literacy model specify that informal literacy activities primarily influence children’s language development, whereas formal literacy activities more directly influence children’s literacy skills (e.g., alphabet knowledge, word reading). Evidence generally supports these hypotheses (e.g., Hood, Conlon, & Andrews, 2008).

Although the home literacy model is supported by empirical evidence, it specifically focuses on literacy-related practices and does not explicitly consider other aspects of the HLE (e.g., conversational interactions with caregivers and siblings that are not contextualized on print materials) that may influence children’s developing language skills. Schmitt, Simpson, and Friend (2011) proposed that the HLE can be defined either conservatively, focusing only on specific reading activities, or broadly, considering specific reading activities alongside other important contextual factors, such as frequency of child-caregiver interactions or maternal education. Evidence that quantity and quality of language input (not specifically focused on literacy activities) influences language development (e.g., Weisleder & Fernald, 2013) provides support for a broader conceptualization of the HLE that goes beyond examining specific literacy-related activities. However, there is also some evidence to suggest that the more active, literacy focused HLE activities play the largest role in children’s language development (e.g., Burgess, Hecht, & Lonigan, 2002; Umek, Podlesek, & Fekonja, 2005). Nevertheless, examining influences of the HLE on language development through a broader lens that includes non-literacy-focused language exposure may be particularly important for understanding how the HLE influences language development for DLLs. Some aspects of the HLE (e.g., quantity of exposure to English speech) may be more strongly associated with language development for DLLs than for monolingual children because language input is distributed across multiple languages for children growing up in dual-language contexts. Therefore, for the purposes of this study, we utilized a broad conceptualization of the HLE that included DLLs’ exposure to English and Spanish language input, English and Spanish literacy activities (e.g., shared storybook reading), and family beliefs/reading habits (e.g., parental reading skills and enjoyment of reading in Spanish and English).

Associations between the HLE and Language Development among DLLs

Prior research suggests that the HLE does play a significant, and potentially causal, role in the development of language skills for preschool-age DLLs. Specifically, parental involvement in literacy activities and home literacy resources are positively associated with young DLLs’ vocabulary and early literacy outcomes (Farver, Xu, Eppe, & Lonigan, 2006; Farver, Xu, Lonigan, & Eppe, 2012). Furthermore, quality of language input appears to play a role in vocabulary development, as prior research indicates that the frequency with which mothers ask questions during shared book reading is significantly related to DLLs’ English and Spanish vocabulary development (Quiroz et al., 2010). Consistent with the findings of Quiroz et al., experimental evidence indicates that interventions focused on improving the HLE (e.g., shared book reading) are effective for improving language outcomes among Spanish-speaking DLLs (e.g., Fitton, McIlraith, & Wood, 2018; Hammer & Sawyer, 2016).

Some additional evidence indicates that the HLE continues to influence development of vocabulary knowledge among DLLs into the elementary years. For example, Uchikoshi (2006) reported that exposure to books and viewing educational television at home were significantly related to level of English vocabulary at the beginning of kindergarten. Additionally, Dixon et al. (2012) reported that effects of language exposure on vocabulary development were language-specific among kindergarten DLLs in Singapore. Similarly, Duursma et al. (2007) reported that parental and sibling support for English were the most consistent predictors of English and Spanish vocabulary knowledge among fifth grade DLLs, such that greater support for English was associated with more English and less Spanish vocabulary knowledge. Other significant predictors of English vocabulary knowledge included personal supports for English literacy development at home (e.g., helping with homework in English), reading or looking at books in English, and telling stories to children in English. In summary, evidence indicates that the HLE is associated with DLLs’ language development throughout early childhood, and several studies suggest these associations are language-specific (e.g., Dixon et al., 2012; Farver et al., 2013; Prevoo et al., 2014; Quiroz et al., 2010; Scheele et al., 2010).

Although there is substantial evidence that the HLE is important for DLLs’ developing language skills throughout early childhood and the elementary years, few studies have examined which specific aspects of a broadly-defined HLE influence growth in Spanish and English vocabulary knowledge for Spanish-speaking DLLs. One longitudinal study of growth in vocabulary knowledge among Spanish-speaking DLLs from 4 to 12 years of age (Mancilla-Martinez & Lesaux, 2011) indicated that children who were exposed to mostly English at home had greater initial English vocabulary than did children who were exposed to mostly Spanish or English and Spanish equally. However, as children moved through the elementary school years (95% of children were in English-only classrooms), children who were exposed to mostly Spanish or English and Spanish equally at home demonstrated substantially faster rates of growth in English vocabulary than did children who were exposed to mostly English at home, resulting in no gaps in English vocabulary by 12 years of age across groups. In contrast, children who were exposed to mostly Spanish at home had the highest level of Spanish vocabulary, followed by children who were exposed to English and Spanish equally and children who were exposed to mostly English. Gaps in Spanish vocabulary knowledge did not change from 4.5 to 12 years of age. Results of this study were consistent with theoretical predictions that the influence of the HLE is language-specific for young DLLs, and indicate that the early childhood years, prior to school entry, represent a key developmental period during which the HLE influences children’s language development. Given evidence of large 5^th grade achievement gaps between monolingual children and DLLs who start kindergarten with limited English proficiency, but no gaps for DLLs who begin kindergarten proficient in English (Kieffer, 2008), more research is needed to identify the aspects of the HLE that promote successful acquisition of Spanish and English language skills in early childhood.

Current Study

The purpose of this study was to examine which aspects of the HLE significantly predicted initial level and rate of growth of English and Spanish vocabulary among young Spanish-speaking DLLs. To do so, we first evaluated the nature of growth of English and Spanish vocabulary separately. Then, based on prior theory and evidence of how the HLE influences language development for DLLs, we examined which aspects of the HLE (i.e., language exposure, reading exposure, family reading habits) were predictive of initial level and rate of growth of English and Spanish vocabulary. Finally, we evaluated models that included all aspects of the HLE together to determine the unique influence of various aspects of the HLE on vocabulary development. All HLE predictors were examined when controlling for child age, socioeconomic status (as indexed by maternal education and annual family income), and initial vocabulary in the language not being examined (e.g., controlling for initial English vocabulary when examining growth in Spanish vocabulary). Given prior theory and research, it was anticipated that input in Spanish and English would be predictive of growth in vocabulary knowledge but that these effects would be language specific. Additionally, we expected that exposure to literacy materials and activities in the home would be associated with greater growth in vocabulary knowledge.

Method

Participants

This research was approved by the Institutional Review Board of Florida State University. Data for this study came from a sample of 944 Spanish-speaking children who were recruited across three consecutive years from different regions across the U.S. for participation in a large preschool curriculum evaluation project. Participating children were followed over the course of two school years. Children were enrolled in 113 classrooms across 83 preschool centers. Most children were assessed during the final year of preschool and in kindergarten the following year; however, a small number of children completed assessments during two years of preschool. Children participating in this study were enrolled in Head Start centers, preschool classrooms in public school districts, and private preschools that were full day preschool centers children attended five days per week. All preschool classrooms participating in this study had at least one primary teacher or assistant teacher who was proficient in Spanish. Although both monolingual English-speaking and Spanish-speaking DLLs participated in the larger curriculum evaluation project, data reported in this study only came from the Spanish-speaking portion of the sample, as the research questions for this study focused on aspects of the home language environment that influence growth in Spanish and English oral language skills. Approximately half of the sample was female (51.6%), and the average age of children at the beginning of the study was 53.77 months (SD = 5.03 months).

Children were identified as Spanish speakers using a combination of parent report, teacher report, assessments of Spanish language proficiency, and brief interactions between children and research assistants to determine children’s preferred conversational language. For all children identified as Spanish speakers by parent or teacher report, research assistants interacted with the child in Spanish. If the child responded to research assistants’ use of Spanish with use of Spanish, the child was determined to be a Spanish speaker and completed assessments in Spanish. Children were recruited from preschool centers in Florida, California, New Mexico, and Kansas. Consequently, children in this sample represented a diverse group of Spanish speakers, including Spanish dialects from Mexico, Cuba, Puerto Rico, the Dominican Republic, and numerous countries in Central and South America. Data on child race and ethnicity indicated that 921 (97.7%) identified as White and 11 (1.2%) identified as Black or African American. Race did not differ across sites, as all sites except Kansas (92.2%) had greater than 97% of participants who identified as White. Individuals identifying as Hispanic comprised 96.4% of the sample. Among children for whom parent-report data were available (80.6% of the sample), 95.0% of children either preferred to speak Spanish or spoke English and Spanish equally. For 77.8% of the sample, Spanish was reported to be the language most spoken at home, and an additional 11.7% reported that Spanish and English were spoken equally at home. Most children were first generation U.S. citizens, with only 5.1% of children born outside the U.S. or Puerto Rico. In contrast, 83.1% of mothers and 86.3% of fathers were born outside the U.S. or in Puerto Rico. Because children were recruited for participation in an intervention study, children from at-risk backgrounds were specifically recruited for inclusion in this study. Consequently, children in this sample primarily came from lower SES backgrounds. Average maternal education in this sample was having completed 9^th-12^th grade, and average annual family income was between $20,000 and $25,000.

Measures

Vocabulary Knowledge.

Children completed the Definitional Vocabulary subtest of the Test of Preschool Early Literacy (TOPEL; Lonigan, Wagner, Torgesen, & Rashotte, 2007). This test assesses children’s expressive vocabulary and conceptual knowledge. Each of the 35 items on this subtest has two components, a naming and a definitional component. For the naming component, children view a picture (or group of pictures) and are required to provide one word that describes the concept (or concepts) pictured. For example, for one item, children view a picture of a monkey and are asked what is this? For the definitional components of items, children are asked a follow-up question that requires them to describe the object or a feature of the object. For example, for the item in which children view a picture of a monkey, children are asked where does it live? All answers are scored as incorrect or correct (i.e., 0 or 1). Consequently, children can receive scores ranging from 0 to 70, as children receive one point for each naming component answered correctly and one point for each definitional component answered correctly. If children give an incomplete answer or an answer that is too broad they are prompted to be more specific (e.g., tell me more). If children respond incorrectly to both the naming and definitional components of three consecutive items, testing is discontinued. Internal consistency reliability for this subtest is high (α = .98). Although the TOPEL was normed using a sample that primarily consisted of monolingual English-speaking children, recent evidence indicates it does not demonstrate substantial bias when used with Spanish-speaking populations (Goodrich, Lonigan, & Alfonso, 2019).

Children also completed the Definitional Vocabulary subtest of the Spanish Comprehensive Test of Phonological and Print Processing (Lonigan, Farver, & Eppe, 2002). This test is a Spanish-language adaptation of the Definitional Vocabulary subtest of the TOPEL and is identical to that subtest in structure. Internal consistency reliability for this subtest was high in this sample of children (α = .98). Items on the Definitional Vocabulary subtests were only scored as correct if responded to in the language being assessed. If children responded in the other language (e.g., responded in English to the Spanish Definitional Vocabulary subtest), they were prompted to respond in the correct language once per item. If the child still responded in the incorrect language, the item was scored as incorrect and testing proceeded with the next item.

Home language survey.

Parents of participating children completed a Family Demographics Worksheet as well as a Language and Reading Survey developed for this study that measured information about children’s home language and literacy environment. The scale and anchor values for HLE variables are reported in Table 1. Parents had the option to complete the measure in Spanish or English, depending on the language they were most comfortable with. The Family Demographics Worksheet included information related to the child’s sex, race, ethnicity, and age. Additionally, this survey included information about the language spoken by and with the child, parental age, country of birth, education level, and family income. The Language and Reading Survey included questions that measured information about the language to which children were exposed daily, children’s reading exposure, and family reading habits. Thus, the survey was divided into three separate subscales: Language Exposure, Reading Exposure, and Family Reading Habits.

Table 1.

Anchor values corresponding to numeric data for the home language survey.

	Time Spent Watching TV Daily
Scale	0	1	2	3	4	5
Value	No TV	0 to 1 Hours	1 to 2 Hours	2 to 3 Hours	3 to 4 Hours	4+ Hours

	Percent of Time Spanish is Spoken at Home
Scale	0	1	2	3	4
Value	None	1–25%	26–50%	51–75%	76–100%

	Frequency of Conversations/Reading with Child
Scale	0	1	2	3	4	5	6
Value	Never	1x a Week	2x a Week	3x a Week	4x a Week	5x a Week	Almost Daily

	Number of Picture Books in the Home
Scale	0	1	2	3	4	5	6	7	8	9	10
Value	None	1–10	11–20	21–40	41–60	61–80	81–100	101–150	151–200	201–300	300+

	Number of Alphabet Books in the Home
Scale	0	1	2	3	4	5	6
Value	None	1–5	6–10	11–15	16–20	20–30	30+

	Language Read to Child
Scale	1	2	3	4	5
Value	English Only	Mostly English	50/50 English/Spanish	Mostly Spanish	Only Spanish

	Child’s Enjoyment of Reading
Scale	1	2	3	4
Value	A Little	Pretty Much	Very Much	Loves it

	Frequency of Engaging Child in the Story
Scale	0	1	2	3	4
Value	Never	Rarely	Sometimes	Usually	Always

	Frequency of Parent Reading for Fun or Pleasure
Scale	0	1	2	3	4	5	6
Value	Never	Almost Never	Monthly	Twice a Month	Weekly	Every Other Day	Daily

	Parent Reading Skills in English/Spanish
Scale	1	2	3	4
Value	Poor	Adequate	Good	Outstanding

	Importance of Preparing Child for Preschool
Scale	0	1	2	3	4
Value	Not Important	A Little Important	Somewhat Important	Important	Very Important

	Maternal Education
Scale	1	2	3	4	5	6
Value	Elementary to 6th Grade	7th to 8th Grade	9th to 12th Grade	1 to 2 Years College	3 to 4 Years College	College Grad or Higher

	Family Income
Scale	1	2	3	4	5	6	7
Value	5k	10k	15k	20k	25k	30k	31k-40k
Scale	8	9	10	11	12	13	14
Value	41k-50k	51k-75k	76k-100k	101k-125k	126k-150k	151k-175k	175k+

To measure Language Exposure, the survey included items assessing amount of time spent watching television, percent of time Spanish was spoken at home, and how often parents engage in conversations with children. To measure Reading Exposure, the survey included questions regarding the number of books available at home, the frequency with which parents read to children, the language in which parents read to children, how often parents engaged children during storybook reading, and the child’s interest in reading. To measure Family Reading Habits, the survey included questions regarding parental interest in reading, self-report of parental reading skills in Spanish and English, and the importance parents placed on preparing their children for school.

Procedure

Participating children completed assessments at the beginning and end of each school year (four total assessment points). Parents completed the Family Demographics Worksheet and Language and Reading Survey at the beginning of Year 1 of the study. Trained bilingual research assistants administered all direct child assessments individually in a quiet area of children’s preschool centers or schools. To prevent fatigue, individual assessment sessions lasted no longer than 30 minutes. Spanish and English assessments were administered on separate days, usually no more than one or two weeks apart, and order of assessment of Spanish and English subtests varied across participants.

Results

Descriptive Statistics

Average raw scores for the English and Spanish Definitional Vocabulary assessments at each time point are reported in the upper panel of Table 2. Preliminary analysis indicated that Initial Spanish vocabulary scores were slightly higher than were English vocabulary scores at Time 1, t(925) = 4.15, p < .001; however, by the end of the first year of the study children’s English vocabulary scores were significantly higher than were their Spanish vocabulary scores, t(692) = 6.72, p < .001, and the gap between English and Spanish vocabulary scores widened over the course of Year 2. Children’s English and Spanish vocabulary knowledge were not significantly correlated at any time point (−.02 < r < .08). Standard scores are not available for the Spanish Definitional Vocabulary subtest, and standard scores on the English Definitional Vocabulary subtest indicated that children’s English vocabulary at the beginning of Year 1 was well below average (Time 1 M = 72.01, SD = 18.86). As children progressed through preschool and kindergarten, average English vocabulary increased to the low-average range (Time 4 M = 89.29, SD = 14.52).

Table 2.

Descriptive statistics for child age, vocabulary knowledge, and aspects of the home language and literacy environment

	T1 Mean (SD)	T2 Mean (SD)	T3 Mean (SD)	T4 Mean (SD)
Child Age	53.77 (5.03)	60.75 (4.94)	66.22 (5.16)	72.58 (4.81)
English Vocabulary	21.90 (19.32)	40.73 (17.05)	46.45 (14.34)	53.92 (10.62)
Spanish Vocabulary	25.38 (15.38)	33.70 (17.96)	33.88 (19.12)	37.14 (18.60)
	Mean	SD	Min	Max
Time Spent Watching TV Daily	2.71	1.17	0	5
% Spanish Spoken in Home	2.88	1.29	0	4
Frequency of Conversations with Child	5.82	.75	0	6
Number of Picture Books in Home	2.52	1.75	0	10
Number of Alphabet Books in Home	1.74	1.35	0	6
Frequency of Reading to Child	3.75	1.89	0	6
Language Read to Child	3.23	1.25	1	5
Child’s Enjoyment of Reading	3.29	.87	1	4
Frequency of Engaging Child in Story	2.62	1.14	0	4
Parent Reads for Fun or Pleasure	4.26	1.69	0	6
Parent English Reading Skills	2.35	1.03	1	4
Parent Spanish Reading Skills	3.16	.79	1	4
Importance of Preparing Child for Preschool	3.62	.67	0	4
Maternal Education	3.34	1.57	1	6
Family Income	4.52	2.41	1	14

Note. Maximum score on English and Spanish vocabulary assessments was 70.

Because children in this sample came from diverse geographic regions across the U.S., we examined whether children’s English and Spanish vocabulary knowledge differed across region. Child age significantly differed across geographic region, F(3,932) = 57.63, p < .001, with children from California being approximately 4–5 months younger than children from Kansas, New Mexico, and Florida at Time 1 (all ps < .001; no other differences across region). Therefore, we examined differences in English vocabulary using standard scores. Because standard scores for Spanish vocabulary were not available, we examined differences in Spanish vocabulary knowledge across geographic region using analysis of covariance, controlling for child age. Univariate analysis of variance indicated that standard scores for English vocabulary at Time 1 did not differ across geographic region, F(3,932) = 1.90, p = .13. In contrast, after controlling for child age, there were significant differences in Spanish vocabulary knowledge across geographic region, F(3,923) = 12.99, p < .001. Specifically, children in New Mexico (adjusted M = 33.20) had significantly greater Spanish vocabulary knowledge than did children in all other states (all ps < .001; California adjusted M = 24.43, Florida adjusted M = 26.01, Kansas adjusted M = 19.13) and children in Kansas had significantly lower Spanish vocabulary knowledge than children in all other states (all ps < .01). The difference in Spanish vocabulary knowledge between children in Florida and California was not statistically significant (p = .22).

Means, standard deviations, and the possible range of values for HLE variables included in growth models are reported in the lower panel of Table 2. The modal response for parent report of time spent watching television daily was 2–3 hours, and it was reported that children viewed educational programming for most of the time spent watching television. Parents reported having conversations with children daily and reading with children approximately four times per week. The modal response for number of picture books available in the home was 1–10, followed by 11–20 and 21–40. Parents reported having stronger reading skills in Spanish than in English, and data on language use at home indicated that the modal response for percent of time Spanish was spoken at home was 76–100% (47.2% of the sample), followed by 26–50% (18.4% of the sample) and 51–75% (17.3% of the sample). Language use at home differed by geographic region, F(3,729) = 9.48, p < .001. Specifically, families in the Florida sample reported using Spanish at home significantly more often than did families in other regions (all ps < .05). Families in California spoke Spanish at home more often than did families in New Mexico (p < .05). No other group differences were statistically significant. In contrast to language use at home, the modal response was that parents read to children equally in Spanish and English (32.7% of the sample). An additional 41.5% of the sample indicated that home reading activities occurred either mostly (22.0%) or only (19.5%) in Spanish, whereas 25.8% of the sample indicated that home reading activities occurred mostly (14.1%) or only (11.7%) in English. Zero-order correlations between Spanish and English vocabulary knowledge and variables from the Language and Reading Survey included in growth models are reported in Table 3.

Table 3.

Correlations between the home language and literacy environment and children’s initial Spanish and English language skills.

	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17
1. TV	.06	.09*	.06	−.02	−.06	−.04	.02	−.05	−.02	.09*	.00	.04	.12**	.11*	.10**	−.05
2. % Spn		.05	−.02	−.01	.03	.20**	.08*	.02	.06	−.11**	.22**	.04	.11**	.09+	−.01	.09*
3. Conv Freq			.10**	.10*	.17**	−.05	.14**	.11**	.13**	.05	.06	.19**	.15**	.06	.04	.01
4. Picture Book				.62**	.35**	−.24**	.23**	.14**	.22**	.20**	.06	.13**	.31**	.22**	.26**	−.02
5. ABC Book					.26**	−.15**	.24**	.27**	.15**	.15**	.10*	.12**	.14**	.10*	.15**	−.01
6. Read Freq						−.09*	.36**	.32**	.26**	.05	.05	.17**	.12**	.13**	.14**	.02
7. Lng Read							−.05	.03	−.08*	−.58**	.15**	−.04	−.30**	−.26**	.35**	.19**
8. Enjoy Read								.29**	.21**	.10*	.03	.26**	.18**	.03	.08*	.05
9. Engage Story									.18**	.06	.11**	.19**	.02	−.07	.02	.01
10. Par Eng Read										.16**	.18**	.17**	.17**	.08+	.09*	.04
11. Par Eng Read											.06	.08*	.44**	.34**	.26**	−.06
12. Par Spn Read												.09*	.23**	.18**	−.02	.13**
13. Pre-K Importance													.16**	.03	.02	.04
14. Maternal Ed														.46**	.28**	.09*
15. Income															.26**	.03
16. Eng Voc T1																−.02
17. Spn Voc T1

Note. TV = Average number of hours spent watching TV per day. % Spn = Percent of time Spanish is spoken at home. Conv. Freq. = Frequency with which parents have conversations with children. Picture Book = Number of picture books in the home. ABC Book = Number of alphabet books in the home. Read Freq. = Frequency with which parents read to children. Enjoy Read = Child enjoys reading. Engage Story = Frequency with which parents engage children in stories. Par. Enj. Read = Parents’ self-reported enjoyment of reading. Par Eng Read = Parental reading skills in English. Par Spn Read = Parental reading skills in Spanish. Pre-K Importance = Parental beliefs about the importance of preparing children for preschool. Maternal Ed. = Maternal education level. Income = Annual family income. Eng Voc T1 = English definitional vocabulary knowledge at Time 1. Spn Voc T1 = Spanish definitional vocabulary knowledge at Time 1.

^**p < .001.

^*p < .05.

⁺p < .10.

Unconditional Latent Growth Models

Prior to conducting tests of the primary research questions, latent growth curve models were fit to Spanish and English vocabulary data using Mplus Version 7.4 (Muthén & Muthén, 1998–2012). Any missing data were handled using full-information maximum likelihood estimation procedures, and all growth models were fit using the sandwich estimator to adjust standard errors in accordance with the nested structure of the data (i.e., children nested within preschool centers). For all growth models, the intercept was centered at Time 1. Growth models were evaluated for Spanish and English vocabulary separately to determine the nature of growth in these skills over the course of Years 1 and 2. Fit statistics for these models are reported in Table 4, and patterns of growth of Spanish and English vocabulary are plotted in Figure 1. Both models provided good fit to the data. For Spanish vocabulary, a model that only included a linear growth term provided the best fit to the data. There were significant individual differences in both initial level and rate of growth of Spanish vocabulary. In contrast, for English vocabulary, inclusion of both linear and quadratic growth terms provided the best fit to the data. The variance for the intercept and quadratic growth terms were statistically significant, indicating that there were significant individual differences in initial level of English vocabulary and in the rate of acceleration/deceleration of growth; however, there were no individual differences in linear growth rate of English vocabulary. The variance of the linear growth term was fixed to zero, as it was negative and not statistically significant.

Table 4.

Model fit statistics for latent growth models of Spanish and English vocabulary

	Unconditional Models
	Chi-Square	RMSEA	CFI	TLI	SRMR
English Linear	59.26***	.11	.85	.83	.11
English Quadratic	18.72***	.06	.96	.94	.07
Spanish Linear	24.05***	.06	.95	.94	.04
Spanish Quadratic	19.77***	.14	.95	.70	.03
	Conditional Models – Spanish
Model 1	31.24**	.04	.98	.96	.03
Model 2	35.71*	.03	.98	.97	.02
Model 3	57.83***	.04	.97	.95	.02
Model 4	43.78**	.03	.98	.96	.02
Final Model	49.79***	.04	.97	.95	.02
	Conditional Models – English
Model 1	18.14**	.05	.99	.96	.04
Model 2	22.69**	.04	.99	.96	.03
Model 3	28.51**	.04	.99	.96	.02
Model 4	26.60**	.04	.99	.95	.03
Final Model	23.10**	.04	.99	.96	.03

Note. N = 942. RMSEA = Root mean square error of approximation. CFI = Comparative fit index. TLI = Tucker-Lewis Index. SRMR = Standardized root mean square residual.

^***p < .001.

^**p < .01

^*p < .05.

Figure 1.

Estimated growth trajectories for English and Spanish vocabulary beginning at Time 1 (0 months) and ending at Time 4 (18 months).

Conditional Growth Models

To examine the effects of the HLE on level and rate of growth of Spanish and English vocabulary, a series of growth models was tested. Model 1 included child age and vocabulary knowledge in the other language at Time 1, as well as maternal education and family income (as indicators of SES) as predictors. These control variables were included in all subsequent models. Because there were many potential variables from the Family Demographics Worksheet and Language and Reading Survey, predictors were added from each of the subscales (i.e., Language Exposure, Reading Exposure, and Family Reading Habits) separately. Model 2 tested the relations between Language Exposure items and initial level and rate of growth of vocabulary. Model 3 tested the relations between Reading Exposure items and initial level and rate of growth of vocabulary. Model 4 tested the relations between Family Reading Habits items and initial level and rate of growth of vocabulary. Models 2 through 4 also included all predictors from Model 1. Finally, a conditional model that combined the significant predictor variables from Models 1 through 4 was examined. For all models, the effect of initial level of vocabulary on rate of growth was estimated. Results of these models are described separately for English and Spanish vocabulary below.

Growth in Spanish vocabulary.

Results for Spanish vocabulary are reported in Table 5. In Model 1, child age and maternal education were positively related to initial level of Spanish vocabulary, and number of words known in English at Time 1 was negatively related to initial level of Spanish vocabulary. This pattern of results indicated that older children, children whose mothers had higher levels of education, and children who knew fewer words in English at Time 1 had higher Spanish vocabularies at Time 1. Additionally, maternal education was negatively related to rate of growth in Spanish vocabulary, indicating that although children whose mothers had higher levels of education had higher initial Spanish vocabulary, these children experienced less growth in Spanish vocabulary over time.

Table 5.

Standardized parameter estimates for models predicting growth in Spanish vocabulary from various aspects of the home language and literacy environment

	Intercept	Linear	Quadratic
Model 1
Intercept		.11	--
Age	.34***	−.10	--
English Vocabulary	−.64***	.21	--
Maternal Education	.17***	−.24*	--
Family Income	.04	−.08	--
Model 2
Intercept		.11	--
Age	.34***	−.11	--
English Vocabulary	−.64***	.25	--
Maternal Education	.17***	−.28**	--
Family Income	.04	−.09	--
Time Spent Watching TV Daily	.00	−.06	--
% Spanish Spoken in Home	.02	.27***	--
Frequency of Conversations with Child	.00	.00	--
Model 3
Intercept		.14	--
Age	.35***	−.13	--
English Vocabulary	−.63***	.36+	--
Maternal Education	.16**	−.17	--
Family Income	.05	−.03	--
Number of Picture Books in Home	.07	−.08	--
Number of Alphabet Books in Home	.00	.00	--
Frequency of Reading to Child	.05	.09	--
Language Read to Child	.10*	.33***	--
Child’s Enjoyment of Reading	.04	−.03	--
Frequency of Engaging Child in Story	.03	−.08	--
Model 4
Intercept		.09	--
Age	.34***	−.08	--
English Vocabulary	−.62***	.26	--
Maternal Education	.17***	−.22*	--
Family Income	.05	−.05	--
Parent Reads for Fun or Pleasure	.06	.07	--
Parent English Reading Skills	−.07	−.25*	--
Parent Spanish Reading Skills	.04	.24***	--
Importance of Preparing Child for Preschool	.01	−.02	--

Note.

^***p < .001.

^**p < .01.

^*p < .05.

⁺p < .10.

In Model 2, percent of time Spanish was spoken at home was positively related to rate of growth of Spanish vocabulary. This pattern of results indicated that children who were exposed to more Spanish at home experienced faster rates of growth in Spanish vocabulary than did children who were exposed to less Spanish at home. No other effects were significant; therefore, the amount of time spent watching television and the frequency with which parents had conversations with children were not included in subsequent models.

In Model 3, the language in which parents read to their child was positively related to initial level of and rate of growth of Spanish vocabulary knowledge. This finding indicated that children whose parents reported reading to their children more in Spanish than in English began preschool with greater Spanish vocabulary knowledge and had faster rates of growth of Spanish vocabulary across the preschool and kindergarten years. No other effects were statistically significant; therefore, number of picture and alphabet books in the home, frequency with which parents read to children, children’s interest in reading, and the frequency with which parents engage children in the stories they are reading were not included in subsequent models.

In Model 4, self-reported parental reading skills in Spanish were positively related to initial level and rate of growth of Spanish vocabulary. Additionally, self-reported parental reading skills in English were negatively related to rate of growth of Spanish vocabulary. This pattern of results indicated that children of parents with stronger Spanish reading skills experienced more growth in Spanish vocabulary whereas children of parents with stronger English reading skills experienced less growth in Spanish vocabulary. No other effects were statistically significant; therefore, parental interest in reading and importance placed on preparing children for school were not included in subsequent models.

When all significant predictor variables from Models 1 through 4 were included in the final model (see leftmost columns of Table 6), child age and level of maternal education were positively related to initial level of Spanish vocabulary. Additionally, number of words known in English at Time 1 was negatively related to initial level of Spanish vocabulary. Percent of time Spanish was spoken at home, language in which parents read to children, and self-reported parental reading skills in Spanish were positively related to rate of growth of Spanish vocabulary. Level of maternal education was negatively related to rate of growth of Spanish vocabulary. This pattern of results was consistent with the patterns observed in Models 1 through 4.

Table 6.

Standardized coefficients for final conditional models examining the unique relations between aspects of the home language and literacy environment and initial level and rate of growth of English and Spanish vocabulary

	Spanish Vocabulary		English Vocabulary
Predictor Variable	Intercept	Linear	Intercept	Linear	Quadratic
Covariates
Intercept		.07		−.30***
Age	.33***	−.09	.26***	.10	−.14
Spanish Vocabulary			−.70***	.14	−.02
English Vocabulary	−.60***	.30+
Maternal Education	.17***	−.24*	.11**	−.03	.06
Family Income	.06	−.05	.07+	.06	−.07
Language Exposure
% Spanish Spoken in Home	.00	.20**
Reading Exposure
Picture Books in Home			.08**	.17**	−.20*
Frequency of Shared Reading			.08*	−.08	.08
Language Read to Child	.08	.16*
Family Reading Habits
Parent Interest in Reading			.02	.01	−.02
Reading Skill in English	−.01	−.14
Reading Skill in Spanish	.20	.20***

Note. Spanish Vocabulary = Spanish vocabulary at Time 1. English Vocabulary = English vocabulary at Time 1. TV = amount of time spent watching TV each day. % Spanish = Percent of time Spanish is spoken in the home.

^***p < .001.

^**p < .01.

^*p < .05.

⁺p < .10.

Growth in English vocabulary.

Results of conditional growth models for English vocabulary are reported in Table 7¹. In Model 1, child age, maternal education, family income, and number of words known in Spanish at Time 1 were significant predictors of initial level of English vocabulary, such that older children, children whose mothers had higher levels of education, children from higher income families, and children who knew fewer words in Spanish at Time 1 had higher English vocabularies at Time 1. Only initial level of English vocabulary at Time 1 significantly predicted linear rate of growth, such that children who had higher initial English vocabulary had slower rates of growth. No other effects were statistically significant.

Table 7.

Standardized parameter estimates for models predicting growth in English vocabulary from various aspects of the home language and literacy environment

	Intercept	Linear	Quadratic
Model 1
Intercept		−.31***
Age	.25***	.09	−.12
Spanish Vocabulary	−.71***	.12	.00
Maternal Education	.14***	.02	.00
Family Income	.09*	.07	−.08
Model 2
Intercept		−.31***
Age	.25***	.09	−.12
Spanish Vocabulary	−.71***	.12	.00
Maternal Education	.14***	.02	−.01
Family Income	.08*	.07	−.08
Time Spent Watching TV Daily	.04	.03	−.01
% Spanish Spoken in Home	−.04	.03	−.03
Frequency of Conversations with Child	.01	−.06	.08
Model 3
Intercept		−.30***
Age	.26***	.10	−.13
Spanish Vocabulary	−.68***	.18+	−.05
Maternal Education	.11**	−.05	.08
Family Income	.06	.05	−.05
Number of Picture Books in Home	.08*	.13+	−.15+
Number of Alphabet Books in Home	−.01	.05	−.06
Frequency of Reading to Child	.07*	−.11	.10
Language Read to Child	−.06+	−.11	.12
Child’s Enjoyment of Reading	.01	.07	−.09
Frequency of Engaging Child in Story	.01	−.03	.04
Model 4
Intercept		−.30***
Age	.25***	.09	−.13
Spanish Vocabulary	−.71***	.13	−.01
Maternal Education	.13***	.03	−.03
Family Income	.09*	.08	−.10
Parent Reads for Fun or Pleasure	.06*	.04	−.07
Parent English Reading Skills	.02	−.03	.06
Parent Spanish Reading Skills	−.04	−.07	.10
Importance of Preparing Child for Preschool	−.01	−.05	.09

Note.

^***p < .001.

^**p < .01.

^*p < .05.

⁺p < .10.

In Model 2, no effects were statistically significant. Therefore, time spent watching TV daily, percent of time Spanish was spoken at home and frequency with which parents had conversations with children were not included in subsequent models.

In Model 3, the number of picture books in the home and the frequency with which parents read to their children were positively related to initial level of English vocabulary, indicating that children from homes with greater numbers of picture books and children whose parents read to them more often had higher initial English vocabulary knowledge. No other effects were statistically significant. Therefore, number of alphabet books in the home, the language in which parents read to their child, children’s interest in reading, and the frequency with which parents engaged children in stories they read together were not included in subsequent models.

In Model 4, parental interest in reading was positively related to initial level of English vocabulary, indicating that children whose parents read more often for fun or pleasure had higher initial levels of English vocabulary. No other effects were statistically significant. Therefore, parental reading skills in English and Spanish and the importance parents place on preparing their children for school were not included in subsequent models.

When all significant predictor variables from Models 1 through 4 were included in the final model (see rightmost columns of Table 6), child age, number of picture books in the home, frequency with which parents read to children, and maternal education were positively related to initial level of English vocabulary. Number of words known in Spanish at Time 1 was negatively related to initial level of English vocabulary. Initial English oral language was negatively related to linear rate of growth. Number of picture books in the home was positively related to linear rate of growth of English vocabulary and negatively related to quadratic rate of growth of English vocabulary. This pattern of results was consistent with the patterns observed in Models 1 through 4.

Discussion

The purpose of this study was to evaluate how aspects of the HLE predicted Spanish-speaking DLLs’ initial level and rate of growth in English and Spanish vocabulary. Partially consistent with our hypotheses, relative English and Spanish input, as measured by the percent of time Spanish was spoken at home and the language in which parents read to children, was associated with growth in Spanish but not English vocabulary knowledge. In contrast, exposure to literacy-related activities at home, such as number of books in the home and the frequency with which parents read to children, was associated with English but not Spanish vocabulary outcomes. These findings are somewhat consistent with prior theory and research indicating that the HLE is significantly related to development of language skills among monolingual children (e.g., Payne et al., 1994) and DLLs (Scheele et al., 2010; Parra et al., 2011) and that influences of the HLE on children’s language skills are language specific (e.g., Farver et al., 2013; Parra et al., 2011; Scheele et al., 2010). Overall, these results provide support for a broadly defined HLE (Schmitt et al., 2011), as language experiences that were and were not focused on reading promoted children’s language development in Spanish and English.

Language-Specificity of Associations between Vocabulary and the HLE

Prior evidence suggests that influences of the HLE on DLLs’ language and literacy skills are language specific (e.g., Dixon et al., 2012; Farver et al., 2013; Prevoo et al., 2014; Quiroz et al., 2010; Scheele et al., 2010). In other words, language and literacy experiences that primarily occur in Spanish should promote development of Spanish language skills (and experiences in English should promote English language skills). The results of our study were somewhat consistent with language-specific effects of the HLE. Children who had greater relative exposure to Spanish at home (compared to English) experienced more growth in Spanish vocabulary. Additionally, children of parents who reported reading to children mostly in Spanish (and who reported greater parental reading skill in Spanish) experienced more growth in Spanish vocabulary. Together, these results indicate that exposure to language and literacy experiences in Spanish positively influences development of Spanish language skills for DLLs in the U.S.

In contrast to results for Spanish vocabulary outcomes, exposure to English was not directly associated with English vocabulary development. It is possible that this is due to differences across children’s Spanish- and English-language environments. According to ecological systems theory (Bronfenbrenner, 2005), development is directly influenced by the interactions children have within their immediate environments (i.e., microsystems). For most DLLs in this study, the two primary microsystems were the home and preschool environments. Given that all children in this study were enrolled in preschool centers, any observed effects of the HLE represent effects on language development beyond the effects of preschool. Many children in this project were enrolled in Head Start programs, which place a heavy emphasis on promoting English language development (U.S. Department of Health and Human Services, 2020). All preschool classrooms participating in this study had at least one primary teacher or assistant teacher who was proficient in Spanish. However, if the primary language of classroom instruction was English (or if use of Spanish in the classroom was relatively homogenous), variation in Spanish exposure at home would primarily drive Spanish language development. Conversely, English exposure at home may not be sufficient to promote English language development, beyond the influence of English language exposure children receive at preschool. This hypothesis is supported by prior evidence that input from native speakers is more important for promoting language development than input from non-native speakers, as well as data indicating that most parents of children in this study were born outside the U.S. and may not have been native English speakers.

Similarly, exposure to children’s books and shared-reading experiences at home influenced development of English but not Spanish vocabulary. Although our data indicated that reading experiences at home were relatively balanced across Spanish and English, the modal response for time spent speaking Spanish at home was 76–100%. Therefore, reading experiences in English may have had unique opportunities to influence English but not Spanish vocabulary, given the less spoken language input in English for DLLs in this sample. Children’s books contain more complex language than is typically used in everyday oral language interactions (Montag, 2019), and evidence indicates that quality of language input plays a significant role in the development of children’s language skills (Weisleder & Fernald, 2013). Given the focus on development of English language skills in Head Start programs, children exposed to English in such programs may need exposure to more complex Spanish language through books to promote Spanish vocabulary development beyond the benefits of preschool.

Implications for Research and Practice

It is important to note that the rate of growth of Spanish vocabulary was substantially less than was rate of growth for English vocabulary. This pattern of growth is consistent with results of other recent evidence on DLLs’ language development (Hoff et al., 2018) that growth in the majority language (in this case English) outpaces growth in the home language and that the home language is the more vulnerable language. To promote and maintain development of Spanish, it is important for parents and early education providers to provide targeted, high-quality language and literacy activities in Spanish, including exposure to Spanish-language children’s books.

The negative influence of maternal education on Spanish vocabulary development indicates that although children from lower SES backgrounds began the study with lower Spanish vocabularies than did their peers from higher SES backgrounds, Spanish vocabulary knowledge of children from lower SES backgrounds grew faster than did Spanish vocabulary knowledge of children from higher SES backgrounds. This suggests that the language gap between groups of children from higher and lower SES backgrounds narrowed over the course of the study. Children from lower SES backgrounds may have a lower quality of Spanish input at home than do other children; however, attending preschool classrooms with an adult proficient in Spanish may provide these children the high-quality Spanish exposure they need to begin to catch up to their peers. Of course, DLLs come from diverse linguistic backgrounds and home language environments; therefore, many DLLs may enter preschool or kindergarten with well-developed Spanish language proficiency. Future research should explore the extent to which interventions aimed at improving the home language experiences of at-risk DLLs can mitigate the negative effects of SES on children’s early language development in both L1 and L2.

The overall pattern of English vocabulary development was nonlinear, indicating that children’s rate of growth of English vocabulary was slowing down over time. This pattern mirrors development of early language skills among monolingual children (Huttenlocher, Waterfall, Vasilyeva, Vevea, & Hedges, 2010) but may be somewhat delayed due to timing of exposure to Spanish versus English and the fact that DLLs typically have lower single-language vocabulary knowledge than monolingual children (Hoff et al., 2012). Spanish was the primary language spoken at home for most of this sample. Once children received sustained exposure to English in preschool and kindergarten classrooms there may have initially been rapid growth in English vocabulary that then slowed to a more typical developmental trajectory. Consistent with this explanation, prior research indicates that development of vocabulary knowledge among DLL preschoolers outpaces the rate of development of vocabulary among their monolingual peers (Lonigan, Farver, Nakamoto, & Eppe, 2013). This suggests that high-quality language exposure in preschool or kindergarten can help accelerate development of English language skills among children who are not consistently exposed to rich English language environments at home.

Although several variables related to the HLE were significantly associated with initial level of English vocabulary, only number of picture books in the home predicted growth in English vocabulary beyond the influence of initial English vocabulary. Clearly, the complex and rich language interactions shared between parents and children early in life are important for early development of English oral language skills. However, children with greater initial English vocabulary grew less over the preschool and kindergarten years than did other children. Children exposed to more books at home experienced faster growth in English vocabulary but also reached a plateau of English vocabulary sooner than did children exposed to fewer books in the home. These findings suggest that the experiences children have in preschool and kindergarten help narrow the achievement gap between children from different home-language backgrounds, but early education programs may need to provide more systematic, high-quality language environments to promote further development among those DLLs who enter preschool with higher levels of English oral language. Parents can continue to provide more opportunities for growth in English oral language throughout early childhood by ensuring that children have access to high-quality print materials at home to supplement the language and reading exposure they experience at school. Farver et al. (2006) reported that associations between the HLE and children’s language outcomes were mediated by children’s interest in reading. If HLE practices result in increases in young DLLs’ interest in reading, children may be more likely to engage with print materials available in the home, which may in turn result in growth in English oral language skills.

Limitations

This study is one of the first to examine which aspects of the HLE are uniquely predictive of L1 and L2 language growth for DLLs; however, it has several limitations. First, the sample for this study only consisted of Spanish-speaking DLLs growing up in the U.S., the majority of whom came from low SES backgrounds. Patterns of language development and the factors that influence language development may be different for DLLs growing up in different environmental contexts. Additionally, this study relied entirely on a single parent report of the HLE. Future studies should use direct measures of the HLE, such as observations, to gain a more accurate picture of the quantity and quality of language to which DLLs are exposed in L1 and L2. Studies examining change in children’s language skills over time may also consider measuring the HLE at each measurement occasion to capture shifts in the HLE over time, as well as how children’s developing competencies in L1 and L2 may influence the HLE.

Although all participating preschool classrooms included at least one teacher who was proficient in Spanish, future research should incorporate direct measures of individual differences in exposure to Spanish on children’s language development. Some evidence indicates that influences of the HLE on children’s vocabulary skills may differ for children who receive English and Spanish instruction (e.g., Duursma et al., 2007). Consequently, future research should account for variations in language of instruction, as well as the quality of classroom environment/instruction, when examining the influences of the HLE on DLLs’ language development. Similarly, it is possible that English and Spanish language development influence each other. Examination of the co-development of English and Spanish vocabulary resulted in model non-convergence. It is possible that reciprocal relations between English and Spanish language development emerge over long periods of time. Although children completed language assessments at four time points for this study, it represented only two years of language development. Future research should explore the extent to which rate of growth of vocabulary in L1 influences rate of growth of vocabulary in L2, and vice versa. Finally, the first time point represented a point in time at which children were already enrolled in preschool, which may or may not provide high-quality language exposure in one or both languages. It would be informative to know how the HLE influences children’s language development prior to preschool entry when the HLE is the primary means through which children are exposed to language as well how those relations may change once children’s patterns of relative exposure to L1 and L2 shift at preschool entry.

Conclusions

Overall, results of this study indicated that home language and literacy practices significantly influence rate of growth of language skills for young DLLs. Both literacy-focused activities and general language input were associated with vocabulary knowledge, providing support for a broad conceptualization of the HLE for DLLs. These findings highlight the importance of early reading activities in the home for DLLs and reinforce previous findings regarding the language-specificity of the relations between home activities and DLLs’ language development. Future research is needed to describe the mechanisms through which various aspects of the HLE influence child language more clearly. Nevertheless, it is important for parents and practitioners to provide high-quality language and literacy experiences for young DLLs to promote children’s language development.

Highlights.

• Evaluated growth in Spanish and English vocabulary in a large sample of DLLs

• The influence of the home language and literacy environment was language specific

• Language and reading input in Spanish predicted growth in Spanish vocabulary

• Exposure to print at home predicted growth in English vocabulary knowledge

• Language and reading activities at home associated with development of vocabulary