Language exposure outside the home becomes more English-dominant from 30 to 60 months for children from Spanish-speaking homes in the United States

Abstract

Purpose:

The purpose of the present study was to document the out-of-home exposure to English and Spanish experienced by children from Spanish-speaking homes in the United States during the preschool years.

Methodology:

Primary caregivers of 149 children from Spanish-speaking homes in South Florida reported on their children’s language exposure.

Data and analysis:

Descriptive statistics and paired-samples t-tests described and compared children’s exposure to English and Spanish outside the home. Multi-level modeling described trajectories of change and the influence of family characteristics on English and Spanish out-of-home exposure.

Findings:

Children heard more English than Spanish outside of their homes. Grandparents were the primary out-of-home source of exposure to Spanish. Language exposure in preschool and extracurricular activities was primarily English. From 30 to 60 months, English exposure increased, while Spanish exposure decreased. Within this general pattern, there was variability in children’s out-of-home language exposure as a function of parents’ language backgrounds and maternal education.

Originality:

Studies of bilingual children’s language exposure have focused on home language use. The present study shows that out-of-home experiences are a significant source of exposure to societal language (SL) for children from language minority homes.

Implications:

For children in immigrant families, the home and family members outside the home are the primary sources of heritage language exposure. Out-of-home language experience is SL-dominant and increasingly so as children get older, although the degree to which this is the case differs depending on parental characteristics.

Introduction

In order to explain the development of English and Spanish skills in children from Spanish-speaking homes in the United States, it is necessary to identify their sources of language exposure. Most research to date has focused on home language use. However, there is good evidence that out-of-home exposure influences bilingual development. Research with immigrant children in English-speaking Canada has found that out-of-home, and not in home, English exposure predicted the children’s English language growth (Paradis, 2011). Other studies have found that out-of-home sources influence children’s heritage language (HL) skills, as well. For instance, studies have shown that preschool attendance influences both children’s societal language (SL) and HL skills (Biedinger et al., 2015; Bowers & Vasilyeva, 2011; Cabell et al., 2015; Magnuson et al., 2006; Winsler et al., 1999). The aims of the present study were to identify sources of English and Spanish exposure outside the home that children in immigrant Spanish-speaking families in the United States experience and to describe changes in that exposure over the preschool period.

Sources of language input outside the home

Previous studies have identified nonparental family members, preschools, and peers as sources of out-of-home language exposure with effects on the bilingual development of children in immigrant families. For example, among children of Turkish immigrants in Germany, the frequency of Turkish input from relatives was marginally related to the children’s Turkish expressive vocabulary skills at age 3 and 4 years and was significantly related to the children’s Turkish expressive vocabulary at ages 5½ to 6½ years (Biedinger et al., 2015). Children of immigrant families are more likely to be cared for by relatives than to use center-based care in the United States (Brandon, 2002; Radey & Brewster, 2007), and this is likely to be true for children in Spanish-speaking immigrant families in the United States, consistent with the value placed on family in Latino culture.

Language exposure in preschool is the most researched source of out-of-home language exposure for bilingually developing children. Preschool attendance is associated with both SL and HL proficiency. Many studies find that SL skills are more advanced in children who attend preschool (Biedinger et al., 2015; Bowers & Vasilyeva, 2011; Cabell et al., 2015; Magnuson et al., 2006; Winsler et al., 1999). However, a strong bilingual preschool program can benefit HL proficiency without reductions in children’s English proficiency (Winsler et al., 1999). Input from preschool teachers in the United States influences the SL and HL acquisition of students with limited English proficiency (Bowers & Vasilyeva, 2011; Gámez & Levine, 2013). Bowers and Vasilyeva (2011) found that the quantity of teachers’ speech positively related to children’s SL (English) receptive skills. In Biedinger et al. (2015) sample of children of Turkish immigrants, the duration of preschool negatively influenced children’s HL (Turkish) expressive vocabulary at ages 3 through 6½ years.

Effects of peers on language development have also been documented. American Spanish-speaking preschoolers’ English expressive vocabulary was uniquely predicted by their peers’ use of English in the classroom (Palermo et al., 2014). Among children of Turkish immigrants, the number of HL-speaking peers in the preschool classroom affects language acquisition (Biedinger et al., 2015).

Influences of family characteristics on language input outside the home

According to Bronfenbrenner’s bioecological model of development, children are shaped by their immediate experience, but that immediate experience is shaped by larger ecological factors including characteristics of their families (Bronfenbrenner & Morris, 2007). From a different theoretical perspective, that of gene–environment interactions, it has also been pointed out that when children are young, their parents choose their children’s environment (Scarr & McCartney, 1983). Thus, parents make decisions about their children’s out-of-home experiences, and parents’ characteristics may shape the decisions they make. Some immigrant parents emphasize HL maintenance, while others emphasize assimilation through use of the SL. For example, one Mexican-American immigrant parent is reported to have insisted that English be spoken in the home and at school based on his perception that English is superior to Spanish in the United States (Martínez-Roldán & Malave, 2004). Likewise, Curdt-Christiansen (2009) found that Chinese immigrants to Canada had clear beliefs and goals about the value of Chinese, English, and French, and each family’s language policy guided the type of educational support these immigrant parents provided for their children. In the United States, research suggests that socioeconomic status (SES) influences the value mothers place on HL maintenance (Pearson, 2007). Lower SES mothers encourage their children’s English acquisition so their children have the best chance of succeeding in America, particularly in school. Higher SES mothers encourage their children to maintain their HL in addition to acquiring English.

Society’s perception of the HL’s prestige plays an important role in parents’ decisions to emphasize HL maintenance or SL conformity. In the examples used earlier, Martínez-Roldán and Malave’s (2004) Mexican-American and Curdt-Christiansen’s (2009) Chinese-Canadian immigrants developed language decisions, practices, and beliefs that reflected their perception of the HL based on cultural and societal expectations.

Parents’ immigration status is related to the choices they make about their children’s preschool attendance. Multiple studies have found that children whose mothers were born outside of the United States have lower rates of participation in school or center-based preschool programs compared to other children (Karoly & Gonzalez, 2011; Magnuson et al., 2006).

Immigrant parents’ educational attainment is also related to the out-of-home language experiences that they organize for their children. Curdt-Christiansen (2009)’s study of Chinese immigrants in Quebec, Canada, provide support for this argument in that parents’ education background, immigration experiences, and cultural disposition were particularly strong influences on family language policy, which contributed to the type of education parents provided to their children. The people with whom mothers choose to associate impacts with whom their children associate. For example, a mother of low SL proficiency will likely choose friends who use the HL. This choice, based on mothers’ proficiency, influences children’s opportunities for interaction in each language.

In particular, SES has been shown to influence bilingual early childcare experiences (Miller et al., 2014). Miller et al. (2014) report conflicting results that immigrant families with high levels of maternal education were less likely to choose center-based care compared to monolingual families, choosing relative care instead.

Changes in language input outside the home over the preschool years

For families of bilingually developing children, patterns of HL and SL input outside the home may change as children get older. Little research has described the nature of dual language experience across the preschool years. The present study seeks to fill this gap by investigating changes in out-of-home English and Spanish language exposure from 30 to 60 months for children from Spanish-speaking homes and by exploring whether changes over time are influenced by demographic variables.

The present study

Findings indicate that relatives, early childcare and education settings, and peers are important sources of variability in monolingual and bilingual children’s early language experience (Biedinger et al., 2015; Bowers & Vasilyeva, 2011; Cabell et al., 2015; Magnuson et al., 2006; Paradis, 2011; Winsler et al., 1999). The aim of the present study was to describe how grandparents, preschools, and extracurricular activities with peers, function as sources of English and Spanish exposure for preschool aged children from Spanish-speaking immigrant families in a heavily Spanish-speaking region of the United States, South Florida. In addition, the present study asks whether family characteristics influence children’s out-of-home exposure to English and Spanish and whether exposure patterns change over time during the preschool period. The data we use to provide this description and address these questions come from a longitudinal study of Spanish–English bilingually developing children assessed at ages 30, 36, 42, 48, and 60 months.

Method

Participants

The participants were 149 children (75 boys, 74 girls) from Spanish-speaking homes who had at least one immigrant parent from a Spanish-speaking country. Participants were selected from a larger study and met the additional criteria that they had been exposed to Spanish and English from birth.

All families were residents of South Florida, and all children were born in the United States. All children were full term and healthy at birth, with normal hearing. All children were screened for evidence of communicative delay at 30 months. English monolinguals (English input in the home at 30 months was higher than 90%) and exceptional cases (twins, children with language delay or hearing problems, and children who received a significant portion of input from more than two languages) were excluded from this study. Consistent with the heterogeneity of bilingual populations in the United States, family characteristics varied: 72.1% of parent dyads were both native Spanish speakers; 27.9% of parent dyads included one native Spanish and one native English speaker; and 73.0% of mothers had less than a 4-year degree attained in English and 27% of mothers had a 4-year degree or more attained in English. Participants were recruited through advertisements in local magazines and at programs for parents with young children, as well as through word of mouth.

Procedure and instruments

Participating children were longitudinally observed within 6 weeks of turning 30, 36, 42, 48, and 60 months. Measures at each age were collected in three sessions. The first session included an interview with the children’s primary caregiver, conducted by a fully bilingual researcher in the language of the caregiver’s choice. This extensive parent-report interview provided demographic and home language environment information. Caregivers provided estimates of hours per week of English and Spanish exposure with family outside the home, during preschool, and during extracurricular activities including Sunday School, church, structured and unstructured playgroups, library programs, community groups, and sports. For most participants, the data collection visits were conducted in the participants’ homes; approximately 15% preferred conducting the visit at the university laboratories. Parents were given a US$50 gift card at each visit and children were given an age-appropriate gift as compensation for their participation.

Measures and data analysis

We grouped children’s out-of-home experiences into three categories that captured where these 30- to 60-month old children spent their time when they were not at home, based on maternal report. Those categories were: (a) with grandparents; (b) in preschool; and (c) in other extracurricular activities including playgroups, sports activities, and community groups. We then asked how much English and Spanish exposure children received from these sources and how that changed from 30 to 60 months. Finally, we asked whether family variables (parents’ language background and maternal education) influenced either the elevation or slope of changes in these sources of input over time. These questions were pursued in four sets of analyses. The first set of analyses consisted of calculating mean hours per week of English and Spanish exposure that children received outside the home across the preschool period. The second set of analyses used paired samples t-tests comparing hours per week of English to Spanish exposure from sources of out-of-home input to determine the pattern of dual language exposure that children experienced outside the home. The third set of analyses investigated the changes in the hours per week of English and Spanish input from sources outside the home over time, and the final set of analyses tested the effect of parents’ native language (L1) background and maternal education level achieved in the United States on changes in hours per week of English and Spanish exposure from out-of-home sources over time.

Results

Dual language exposure outside the home

To provide an overview of participants’ language experience over the preschool years, children’s mean hours per week of English and Spanish out-of-home exposure were averaged across 30 to 60 months. Paired-samples t-tests revealed that children experienced significantly more English (mean (M) = 17.01, standard deviation (SD) = 10.19) than Spanish (M = 9.14, SD = 8.35) input outside the home, t(148) = 7.452, p < 0.001.

Dual language exposure from sources outside the home

Children’s mean hours per week of English and Spanish exposure from each out-of-home source averaged across 30 to 60 months are presented in Figure 1. Paired-samples t-tests revealed that children experienced significantly more Spanish (M = 4.42, SD = 6.68) than English (M = 0.64, SD = 1.49) input from a grandparent (t(148) = −6.88, p < 0.001) but more English than Spanish input during preschool (English: M = 14.79, SD = 9.84; Spanish: M = 3.95, SD = 5.53; t(148) = 12.45, p < 0.001) and extracurricular activities (English: M = 2.33, SD = 1.99; Spanish: M = 1.10, SD = 1.16; t(148) = 7.18, p < 0.001).

Figure 1.

Mean hours per week of English and Spanish exposure from a grandparent, during extracurricular activities, and during preschool averaged across 30–60 months.

Preschool attendance

Given that input during preschool accounts for roughly 70% of total input outside the home across 30 to 60 months (18.74 mean hours out of 27.22 mean hours), preschool attendance was explored further. The number of children attending preschool increased from 64.4% at 30 months (n = 96) to 85.4% at 60 months (n = 117). Paired-samples t-tests comparing hours attending preschool at 30 months to 60 months revealed that the mean number of hours children spent in preschool increased significantly from 30 months (M = 11.39, SD = 17.53) to 60 months (M = 26.77, SD = 15.25; t(136) = −8.02, p < 0.001). Children who did not attend preschool were included in analyses as having 0 hours per week of preschool attendance. Figure 2 shows the increase in mean number of hours per week during which children attend preschool from 30 to 60 months. These results suggest that a greater number of children are participating in preschool for longer periods of time with increasing age.

Figure 2.

Mean hours per week children attend preschool at 30, 36, 42, 48, and 60 months.

Longitudinal changes in dual language exposure from individual sources outside the home

To ask whether there are longitudinal changes in English and Spanish exposure from sources outside the home, six multi-level models were conducted, one for each language with each input source. For all analyses, children’s age was used to predict growth in the hours per week of language input (English and Spanish, respectively) received from an out-of-home input source (grandparent, preschool, and extracurricular activity, respectively). Results are presented in the base models presented in Tables 1–6 and plotted in Figures 3 and 4.

Table 1.

Effects of parents’ native language (L1) background and maternal education on children’s exposure to English from grandparents between 30 and 60 months (n = 149).

	β	Standard error	p
Model 1 (base model)
Intercept	0.63	0.16	<0.001
Age	0	0.05	0.925
Model 2A (demographic variable model: parents’ L1 background)
Intercept	1.76	0.3	<0.001
Age	−0.17	0.1	0.08
Parents’ L1 background	−1.62	0.36	<0.001
Age × parents’ L1 background	0.25	0.11	0.027
Model 2B (demographic variable model: maternal education)
Intercept	0.52	0.19	0.007
Age	0	0.06	0.998
Maternal education	0.44	0.37	0.233
Age × maternal education	−0.02	0.11	0.871
Goodness of fit	Model 1	Model 2A	Model 2B
−2 log-likelihood model fit index	2787.65	2591.36	2771.12
Akaike information criterion	2795.66	2603.36	2783.12
Bayesian information criterion	2813.33	2629.52	2809.61

Table 6.

Effects of parents’ native language (L1) background and maternal education on children’s exposure to Spanish during extracurricular activities between 30 and 60 months (n = 149).

	β	Standard error	p
Model 1 (base model)
Intercept	1.26	0.14	<0.001
Age	−0.09	0.05	0.045
Model 2 (demographic variable model: parents’ L1 background)
Intercept	0.91	0.28	0.001
Age	−0.01	0.09	0.875
Parents’ L1 background	0.55	0.33	0.099
Age × parents’ L1 background	−0.11	0.11	0.32
Model 2 (demographic variable model: maternal education)
Intercept	1.48	0.16	<0.001
Age	−0.1	0.05	0.061
Maternal education	−0.88	0.31	0.005
Age × maternal education	0.04	0.1	0.672
Goodness of fit	Model 1	Model 2A	Model 2B
−2 log-likelihood model fit index	2558.84	2434.69	2525.99
Akaike information criterion	2566.85	2446.69	2537.99
Bayesian information criterion	2584.37	2472.64	2564.26

Figure 3.

Mean hours per week of English exposure from a grandparent, during preschool, and during extracurricular activities at 30, 36, 42, 48, and 60 months.

Figure 4.

Mean hours per week of Spanish exposure from a grandparent, during preschool, and during extracurricular activities at 30, 36, 42, 48, and 60 months.

English.

The amount of English input from a grandparent did not significantly change over time (Table 1). English input during preschool (Table 2) and extracurricular activities (Table 3) significantly changed over time. Children’s English exposure increased an average of 0.23 hours (approximately 14 minutes) every 6 months during extracurricular activities and an average of 3.21 hours (approximately 3 hours and 13 minutes) every 6 months during preschool.

Table 2.

Effects of parents’ native language (L1) background and maternal education on children’s exposure to English during preschool between 30 and 60 months (n = 149).

	β	Standard error	p
Model 1 (base model)
Intercept	7.53	1.02	<0.001
Age	3.21	0.27	<0.001
Model 2A (demographic variable model: parents’ L1 background)
Intercept	7.21	2.01	<0.001
Age	3.88	0.55	<0.001
Parents’ L1 background	0.37	2.36	0.874
Age × parents’ L1 background	−0.93	0.64	0.145
Model 2B (demographic variable model: maternal education)
Intercept	5.46	1.18	<0.001
Age	3.41	0.32	<0.001
Maternal education	7.82	2.25	0.001
Age × maternal education	−0.75	0.61	0.221
Goodness of fit	Model 1	Model 2A	Model 2B
−2 log-likelihood model fit index	4778.04	4498.22	4743.15
Akaike information criterion	4786.04	4510.22	4755.15
Bayesian information criterion	4803.57	4536.16	4781.42

Table 3.

Effects of parents’ native language (L1) background and maternal education on children’s exposure to English during extracurricular activities between 30 and 60 months (n = 149).

	β	Standard error	p
Model 1 (base model)
Intercept	1.78	0.23	<0.001
Age	0.23	0.07	0.001
Model 2A (demographic variable model: parents’ L1 background)
Intercept	2.67	0.45	<0.001
Age	0.01	0.14	0.953
Parents’ L1 background	−1.17	0.53	0.028
Age × parents’ L1 background	0.31	0.17	0.068
Model 2B (demographic variable model: maternal education)
Intercept	1.61	0.26	<0.001
Age	0.27	0.08	0.001
Maternal education	0.53	0.5	0.296
Age × maternal education	−0.14	0.16	0.372
Goodness of fit	Model 1	Model 2A	Model 2B
−2 log-likelihood model fit index	3099.23	2938.34	3080.62
Akaike information criterion	3107.23	2950.34	3092.62
Bayesian information criterion	3124.76	2976.28	3118.89

Spanish.

Spanish input from a grandparent (Table 4) and during extracurricular activities (Table 6) significantly changed over time. Children’s exposure to Spanish decreased an average of 0.47 hours (about 26 minutes) every 6 months during visits with a grandparent and decreased an average of 0.09 hours (about 5 minutes) every 6 months during extracurricular activities. Spanish input during preschool did not significantly change over time (Table 5).

Table 4.

Effects of parents’ native language (L1) background and maternal education on children’s exposure to Spanish from grandparents between 30 and 60 months (n = 149).

	β	Standard error	p
Model 1 (base model)
Intercept	5.55	0.67	<0.001
Age	−0.47	0.17	0.007
Model 2A (demographic variable model: parents’ L1 background)
Intercept	5.98	1.29	<0.001
Age	−0.63	0.35	0.072
Parents’ L1 background	−0.69	1.51	0.646
Age × parents’ L1 background	0.22	0.41	0.587
Model 2B (demographic variable model: maternal education)
Intercept	4.79	0.78	<0.001
Age	−0.47	0.2	0.02
Maternal education	2.97	1.49	0.046
Age × maternal education	0.01	0.39	0.979
Goodness of fit	Model 1	Model 2A	Model 2B
−2 log-likelihood model fit index	4403.23	4147.69	4371.94
Akaike information criterion	4411.23	4159.69	4383.94
Bayesian information criterion	4428.91	4185.86	4410.42

Table 5.

Effects of parents’ native language (L1) background and maternal education on children’s exposure to Spanish during preschool between 30 and 60 months (n = 149).

	β	Standard error	p
Model 1 (base model)
Intercept	4.5	0.58	<0.001
Age	−0.24	0.16	0.141
Model 2 (demographic variable model: parents’ L1 background)
Intercept	2.55	1.1	0.021
Age	0.15	0.32	0.63
Parents’ L1 background	2.31	1.29	0.076
Age × parents’ L1 background	−0.41	0.37	0.269
Model 2 (demographic variable model: maternal education)
Intercept	4.07	0.68	<0.001
Age	−0.07	0.19	0.727
Maternal education	1.69	1.31	0.198
Age × maternal education	−0.64	0.36	0.076
Goodness of fit	Model 1	Model 2A	Model 2B
−2 log-likelihood model fit index	4133.58	3871.17	4111.85
Akaike information criterion	4141.58	3883.17	4123.85
Bayesian information criterion	4159.11	3909.11	4150.12

Effects of family variables on longitudinal changes in dual language exposure from individual sources outside the home

To ask whether family variables influenced longitudinal changes in children’s language exposure outside the home, parents’ L1 background and maternal education were separately added to each of the multi-level base models described above. Comparisons of model fit were accomplished with Chi-square (χ²) difference tests using the −2 log-likelihood model fit index from each model. The Akaike information criterion and the Bayesian information criterion model fit indices are also reported.

The effect of parents’ L1 background.

Parents’ L1 background was significantly related to the elevation of growth functions for English exposure from grandparents and from extracurricular activities and to the slope of the function for grandparent English exposure. Parents’ L1 background was unrelated to children’s out-of-home exposure to Spanish from any source.

Results indicated a significant improvement in model fit when parents’ L1 background was added to each of the English exposure base models: from a grandparent (χ² (2) = 196.30, p < 0.001), during preschool (χ² (2) = 279.82, p < 0.001), and during extracurricular activities (χ² (2) = 160.90, p < 0.001). There was a significant main effect of parents’ L1 background on children’s English exposure from a grandparent (Table 3) and during extracurricular activities (Table 5), but not during preschool (Table 4). Children with two native Spanish-speaking parents received less English exposure from a grandparent and during extracurricular activities than children with one native English- and one native Spanish-speaking parent.

There was also a significant age × parents’ native background interaction for English exposure from a grandparent (Table 3). As shown in Figure 5, English input from a grandparent increased over time for children with two native Spanish-speaking parents and decreased over time for children with one native English- and one native Spanish-speaking parent.

Figure 5.

Mean hours per week of English exposure from a grandparent from 30 to 60 months, as a function of parents’ native language background.

There was a significant improvement in model fit when parents’ L1 background was added to each of the Spanish exposure base models: from a grandparent (χ² (2) = 255.54, p < 0.001), during preschool (χ² (2) = 262.42, p < 0.001), and during extracurricular activities (χ² (2) = 124.16, p < 0.001). However, there were no significant main or interaction effects of parents’ language background.

The effect of maternal education.

There were significant main effects of maternal education on children’s exposure to English in preschool and a significant main effect of maternal education on the amount of exposure to Spanish from grandparents and extracurricular activities. There were no effects of maternal education on changes over time in English or Spanish exposure from any source. Analyses reveal a significant improvement in model fit when maternal education was added to each of the English exposure base models: from a grandparent (χ² (2) = 16.53, p < 0.001), during preschool (χ² (2) = 34.89, p < 0.001), and during extracurricular activities fit (χ² (2) = 18.61, p < 0.001). There was a significant main effect of maternal education on children’s English exposure during preschool (Table 2). Children with mothers who had a 4-year degree or higher in English tended to receive more English input during preschool than children with mothers who had less than a 4-year degree in English. There were no significant interaction effects. These results suggest that maternal education influences children’s English exposure during preschool.

There was a significant improvement when maternal education was added to the Spanish exposure base models from a grandparent (χ² (2) = 31.29, p < 0.001) and during extracurricular activities (χ² (2) = 32.85, p < 0.001). Model fit did not improve when maternal education was added to the preschool Spanish exposure base model (χ² (2) = 0.31, p = 0.856). A significant main effect of maternal education was observed on children’s Spanish exposure from a grandparent (Table 4) and during extracurricular activities (Table 6). Children with mothers who had a 4-year degree or higher in English received more Spanish input from a grandparent and less Spanish during extracurricular activities than children with mothers who had less than a 4-year degree in English. No interaction effects were significant.

Discussion

Using maternal reports of their Spanish–English bilingual children’s exposure to English and Spanish, the present study described how much exposure to English and to Spanish children in immigrant families in South Florida experience from their grandparents, from preschool attendance, and from extracurricular activities during the preschool years.

We found that throughout the period from 30 to 60 months, children received more English than Spanish exposure during extracurricular activities and preschool. With an average of less than 5 hours per week from 30 to 60 months, extracurricular activities were not a major source of language exposure in either language. During preschool, children received twice the English (M = 8 hours per week) compared to Spanish (M = 4 hours per week) input at 30 months. By 60 months, this gap increased by a factor of 4 such that children received roughly 8 times the English (M = 24 hours per week) compared to Spanish input (M = 3 hours per week). The limited amount of HL exposure that children received during preschool likely comes from their bilingual peers or teachers.

Preschool is the largest single source of out-of-home input for children from 30 to 60 months, and both the percentage of children attending and the number of hours they attend increases over this period. This may explain why overall input outside the home is heavily SL-dominant. Previous research supports the finding that structured events such as preschool reflect a SL-dominant pattern of input. Preschool attendance for children of immigrants positively influenced their SL skills (Magnuson et al., 2006; Winsler et al., 1999). These results may not generalize to areas that have many preschools with a HL or bilingual curriculum. Such schools are rare in South Florida.

Grandparents were a source of Spanish input more than English input over the preschool period. This pattern of input aligns with previous research which found that the quantity of HL input children received from their relatives outside the home influences children’s heritage language acquisition (Biedinger et al., 2015). However, the presence of extended family such as grandparents may not be as ubiquitous as in other immigrant populations. Strict immigration laws and limited financial resources may prohibit multiple generations from relocating to the United States. In these instances, results from the present study may not generalize.

How does children’s exposure to English and Spanish from sources outside the home change during the preschool years?

Results revealed that English exposure during extracurricular activities and preschool increased from 30 to 60 months and Spanish exposure from a grandparent decreased over the same time period. These results suggest that children’s out-of-home language environment becomes more English-dominant over time. The present study is the first to quantify this result outside the home over the preschool period. Combined with the finding that native Spanish-speaking grandparents sometimes speak to their grandchildren, these results reveal that grandparents are not a universal source of HL exposure.

Do family variables influence children’s language exposure outside the home during the preschool years?

Children with a native English-speaking parent experienced more English input from a grandparent and during extracurricular activities than children with two native Spanish-speaking parents. This influence of parents’ L1 background on input is similar to findings of effects of parents’ L1s to children’s home language input. Children with two native Spanish-speaking parents receive less English exposure at home than children with one native Spanish-speaking and one native English-speaking parent (Place & Hoff, 2011).

Parents’ L1 background also affected changes in children’s grandparental input over the preschool years. Children with a native English-speaking parent received increasingly less English from a grandparent from 30 to 60 months, whereas children with two native Spanish-speaking parents received increasingly more English exposure from a grandparent over the same period of time. At 30 months, children with an English-speaking parent received a little over 90 minutes of English input from a grandparent compared to the 15 minutes that children with two Spanish-speaking parents received. By 60 months, these groups nearly converged in that both groups of children were estimated to receive about 45 minutes of English exposure from a grandparent. One explanation for this finding is that families with two native Spanish-speaking parents are making a more concerted effort to incorporate English into their children’s language environment in order to prepare children for the increasing English exposure during organizational activities. In other words, with increasing participation in organizational activities, families with two Spanish-speaking parents may feel societal pressure to speak English and align their family language policy accordingly. This argument is supported by the analyses identifying the L1 of the grandparents seen most often in the present study. The majority of grandparents (an average of 89%) were native Spanish-speakers. For children with two native Spanish-speaking parents, 100% of the grandparents were native Spanish-speakers. These findings reveal that not only are Spanish-speaking grandparents not strictly a source of Spanish, but also that the amount of English these grandparents use with children increases over time. Families with one Spanish-and one English-speaking parent may be more comfortable reducing English exposure from grandparents because children from such families are likely to receive English input elsewhere (i.e. in the home).

The effect of maternal education on children’s language exposure outside the home.

Children of college-educated mothers experienced more English during preschool and less Spanish during extracurricular activities than children of less-educated mothers. They also experienced more Spanish input from a grandparent than children of less-educated mothers. This aligns closely with Miller et al.’s (2014) findings that immigrant families with higher levels of maternal education were more likely to choose relative care. One explanation for these findings may be related to family language policy. Pearson (2007) suggests that family language policies for middle- and high-SES mothers focus on the development of both HL and SL, while family language policies for low-SES mothers focus on SL development. The highest level of societal education was conceptualized as a measure of SES in the present study because SL proficiency limits mothers’ earning power. That is, bilingually developing children with middle- and high-SES mothers (who had a 4-year degree or higher in the SL) experienced more English in preschool and more Spanish from a grandparent, which aligns with a family language policy that emphasizes SL and HL acquisition. In line with a family language policy that emphasizes SL conformity, children with low-SES mothers experienced less Spanish from a grandparent.

Limitations

The input measures in the present study were based on caregivers’ estimations of time spent in each out-of-home source of input and the percentage of English and Spanish children who are exposed to each out-of-home source of input during those experiences. A more reliable measurement would be direct observation.

Another limitation is the limited configurations of parents’ L1 backgrounds. The present study divided participants into only two groups: those with two native Spanish-speaking parents; and those with one native Spanish-speaking and one native English-speaking parent. Previous research has demonstrated that differences in parental categories’ (two HL-speakers; HL-speaking mother, SL-speaking father; SL-speaking mother, HL-speaking father; and two bilingual parents, and two SL-speakers) are related to children’s language output (Gathercole & Thomas, 2009). Although the present study had the data to divide participants into more groups based on parental category, there were not enough participants in each group to make analyses meaningful.

The present study is also limited in that the quality of input from sources outside the home was not analyzed. Measures of grandparents’, preschool teachers’, and peers’ language proficiency would further describe children’s language experiences outside the home. Previous research confirms that the quality of input from relatives influences children’s language skills. Place and Hoff (2011) show that the percentage of English input provided by native speakers contributed to children’s English skills. Likewise, Gathercole and Thomas (2009) found that bilingual parents’ abilities in each language affects the language(s) with which they choose to speak to their children. Unfortunately, the present study did not include a measure of peers’ language proficiency.

Although the standard measure of SES is mother’s highest level of education in any language, the present study used mother’s highest level of education achieved in the United States. The purpose of this was to incorporate an index of SL proficiency into socioeconomic level. This may be considered a limitation because it is a novel conceptualization of SES.

Future directions

The present study did not address the effect of out-of-home input on bilingually developing children’s language acquisition. Given the established effect of home language input on the development of children’s language skills (for review, see Unsworth, 2016), future studies should explore how variation in bilingual input outside the home is related to children’s development of SL and HL proficiency.

The role of family language policy and language prestige on bilingual language experiences outside the home should also be further explored. The present study is limited to the bilingual Spanish–English environment of South Florida, which may not generalize to all dual language environments. Ecological systems theory would suggest family language policy and its effect on children’s language environment would differ as a function of the interaction between environmental layers (Bronfenbrenner, 1979). In other words, bilingual children in South Florida may have a different language experience from bilingual children in a different environment. Cross-cultural studies may provide insight into the variation in bilingual children’s experiences.

Conclusion

Preschool aged children of Spanish-speaking immigrant families in South Florida experience English-dominant language exposure outside their homes. This is true despite the fact that South Florida has a high concentration of Spanish speakers and Spanish is heard in stores, on the radio, and in public announcements. Furthermore, the degree to which children’s out-of-home exposure is English-dominant increased over the period from child age 30 months to 60 months. This study is the first to describe this pattern of language exposure for bilinguals outside the home.

The children’s out-of-home language environments differed as a function of their parents’ L1 backgrounds and maternal education. During the preschool years, children with two native Spanish-speaking parents experienced less English from a grandparent and during extracurricular activities than children with one native Spanish- and one native English-speaking parent. However, the gap in English input from a grandparent nearly closes by 60 months as children with two native Spanish-speaking parents receive more English from a grandparent over time while children with one native Spanish- and one native English-speaking parent receive less English from a grandparent over time. Children with mothers who had a 4-year degree or higher in English received more English exposure and less Spanish exposure during extracurricular activities than those with mothers who had less than a 4-year degree in English. These findings have implications for family language policy. Parents who want their children to develop and maintain the HL need to be purposeful in doing so.

Biographies

Stephanie N Welsh is an undergraduate assistant professor of Psychology. She teaches Introduction to Psychology, Child and Adolescent Development, Research Methods, Statistics, and Neuropsychology. She is an expert in SPSS syntax, and her research interests include bilingual language development, attention, and perception.

Erika Hoff directs the Language Development Lab which comprises several projects aimed at identifying the human abilities and human experiences that contribute to language development. Our main focus is a NICHD-funded longitudinal study of English monolingual and Spanish-English bilingual children from the age of 2 ½ to 10 years. We are looking for evidence of factors in children’s early language experiences and early language development that predict successful oral language and literacy outcomes.