Effects of Migration on the Language and Literacy Practices of Turkish Parents in England

A survey study of the language and literacy practices of first-generation Turkish immigrant families with 3- to 6-year-old children was conducted in England. Information on family socioeconomic background, migration history and language skills of 168 first-generation Turkish parents was collected through structured interviews in Greater London and Northwest England. The study findings suggest that early childhood experiences that are important for the educational attainment of immigrant children may be affected by the family characteristics and the integration experiences of parents. Regression analyses demonstrated that parents from disadvantaged backgrounds engaged in language and literacy activities less often and preferred Turkish as the interaction language at home. Parent’s social integration, measured via parent’s length of residence and English skills, significantly predicted their language use with their children. Low social integration was associated with increased Turkish use, whereas high social integration was associated with more frequent language and literacy activities at home.

attainment of pupils in England has underlined the importance of the home learning environment in early childhood in reducing the educational disparities for ethnic minority pupils from Turkish backgrounds (Strand et al., 2010). To this date, however, child-rearing resources and the risk factors relating to the learning experiences of pupils from Turkish backgrounds have been rarely studied (see also Daglar, Melhuish, & Barnes, 2011). To address this gap in the literature, this study examines the home environment of 3-to 6year-old children in relation to the following research questions: 1. To what extent do the family background characteristics explain the language and literacy resources in the home environment of Turkish children? 2. What are the risk factors for a lack of language and literacy activities in the early years of immigrant children?

Conceptual Framework
The Family Investment Model demonstrates how financial and educational resources of families can influence child-rearing practices drawing on the interactions between social class, parenting behaviours and child development and subsequently contribute to educational inequalities (Conger & Donnellan, 2007). In this model, the researchers posit that financial hardships and lower educational qualification of parents can negatively affect the parenting behaviours through lack of educational materials, knowledge and resources. Poverty and financial stress, for instance, can have a negative effect on the home environment of children underpinning development in early childhood including language. To date, studies have shown that families from disadvantaged backgrounds are less likely to provide stimulating learning activities supporting the language development of children (for a review, see Bus, van Ijzendoorn, & Pellegrini, 1995). Studies on the child-rearing practices of immigrant parents have shown that family socioeconomic background predicts various factors that influence the language learning experiences of children (for a review of literature, see Dixon & Wu, 2014). One of the well-demonstrated effects of family socioeconomic background on the home environment of bilingual children from immigrant families is language use at home. Evidence shows that ethnic minority families from higher socioeconomic backgrounds are more likely to use the national language at home (Cha & Goldenberg, 2015). Studies on Turkishspeaking children in Europe, for instance, shows that family socioeconomic background predicts children's bilingual vocabulary through the mediation of language use at home (Prevoo et al., 2014;Scheele, Leseman, & Mayo, 2010;Westeren, Halberg, Ledesma, Wold, & Oppedal, 2018). In a study of Turkish-Dutch speaking 6-year-old children, Prevoo and her colleagues (2014) demonstrated that mothers from higher socioeconomic backgrounds used Dutch more when communicating and reading with their children. Children of parents, who used Dutch more frequently, then had a larger Dutch vocabulary (Prevoo et al., 2014). For Turkish language skills, however, the researchers found a reverse interaction between mother's education level and home language use. Mothers from higher socioeconomic backgrounds tended to use Turkish less often, which predicted a lower Turkish vocabulary for children (Prevoo et al., 2014). The relevance of maternal behaviour was highlighted in another study of Turkish-German speaking pre-schoolers, which found that mother's education level was a predictor of children's Turkish vocabulary through the home language use, whereas the same did not apply to father's education (Willard, Agache, Jakel, Gluck, & Leyendecker, 2015). This study also reported that mothers with lower education levels used more Turkish at home, which mediated the relationship between mother's education and children's Turkish vocabulary.
The literature demonstrates that family background characteristics not only predict the language environment at home, but also parents' engagement in activities that support the bilingual language development. For example, studies of bilingual children from immigrant families have shown that children from disadvantaged families participated in language and literacy activities less often (Aikens & Barbarin, 2008;Willard et al., 2015) and had fewer literacy resources such as books, media and other written materials (Scheele et al., 2010). The researchers, then, ascribed poor language skills of ethnic minority children partly to a lack of language and literacy resources at home (Howard et al., 2014;Scheele et al., 2010;Willard et al., 2015). For instance, studies of Turkish families in Europe have shown that Turkish parents engaged in language and literacy activities less frequently compared to the native parents in the country, especially if they came from disadvantaged backgrounds (Klein, Biedinger, & Becker, 2014;Leseman, 2000;Prevoo et al., 2014;Scheele et al., 2010;Willard et al., 2015).

Social Integration and Acculturation
Immigration can lead to challenges that may affect parents' child-rearing practices. Many immigrant families experience financial hardships as they start a life in a new country (Ward, Bochner, & Furnham, 2001). Research shows that challenges associated with immigration, such as lack of language proficiency, impermanent residence and unfamiliarity with the legal and healthcare systems can have negative psychological and economic effects on individuals (Luksyte, Spitzmueller, & Rivera-Minaya, 2014). These hardships, in turn, can be detrimental for the home environment (e.g. literacy activities) and positive parenting behaviours (e.g. parental warmth) (Baydar et al., 2014;Chen et al., 2013;Gilbert et al., 2017). Even when parents have the capacity to support their children's learning, language barriers can become an impediment to their participation (Strand et al., 2010). This is congruent with the previously presented evidence showing that immigrant families from disadvantaged backgrounds are less likely to engage in home language and literacy activities that support the language development of their children (Kalia & Reese, 2009;Leseman, 2000;Prevoo et al. 2014).
As immigrant families settle in a new country, they experience changes in their day-to-day lives as a result of contact with a new culture, through readjustments in their lifestyles reflecting social and economic integration into society (Titzmann & Fuligni, 2015). One of the important components of integration is language acculturation (Nguyen & Benet-Martinez, 2013). Language acculturation is especially important for immigrant parents as it can promote their engagement in learning activities which can benefit their parenting efficacy (Costigan & Koryzma, 2011). However, the socioeconomic background of parents may influence language acculturation. Lower educational qualifications and poor language skills can limit parents' access to high-income jobs (Dixon, Wu, & Daraghmeh, 2012) and reduce their interactions with the majority culture (Arriagada, 2005). As a result, families may become more accustomed to speaking their heritage language at home (Zhang, 2008). Studies of bilingual children from Hispanic (e.g. Mexico, Puerto Rico and Cuba) and Asian (e.g. India, China and Japan) families in the US, for instance, have shown that factors such as lower parental educational qualifications and limited financial resources were associated with the increased use of ethnic language at home (Winsler et al., 2014). Similarly, Turkish parents with fewer years of education in Europe reported to use Turkish more frequently when speaking or performing language and literacy activities with their children (Turkish-Dutch: Prevoo et al., 2014;Turkish-German: Willard et al., 2015).
Time spent in the country affects the integration of immigrant families. More time in the country can facilitate social and cultural adjustment and contribute to well-being of immigrant individuals (Markovizky & Samid, 2008). Generally, as immigrant families live in a country longer, they become more proficient in the national language and start communicating in the language more often (Spanish-English: Pease-Alvarez, 2002). This change may also have consequences for the home environment and may influence parents' interactions with their children (Becker, 2010;Quiroz, Snow, & Jing Zhao, 2010;Willard et al., 2015;Westeren et al., 2018), and longer migration history is associated with increased use of the national language at home (e.g. Winsler et al., 2014).
Research on Turkish families in Europe demonstrates similar results (Driessen, Van der Slik, & De Bot, Westeren et al., 2018;Willard et al., 2015). For example, regarding time in the new country, Turkish-German speaking children with first-generation mothers were more likely to have a larger Turkish vocabulary than children of second-and third-generation mothers (Willard et al., 2015). Additionally, a study of Turkish pre-schoolers in Norway has shown that mothers' age at migration predicted children's Norwegian vocabulary (Westeren et al., 2018).
Time in the country and acculturation are linked, and Turkish-German speaking children parent's acculturation experiences (e.g. social network, language skills, German language use and German citizenship) were positively associated with children's cognitive skills, as well as second language proficiency (Becker, 2011). Parent's language use at home, however, mediated the interactions between family acculturation and language outcomes of children. A study of Turkish immigrants in the Netherlands found that the language use at home mediated the effect of mothers' length of residence on children's Dutch vocabulary (Driessen et al., 2002). Willard et al. (2015) similarly reported that increased Turkish use at home mediated the relationship between mother's generational status in the country and children's Turkish vocabulary, whereas the same effect was not found for father's generational status.
Drawing on these issues experienced by immigrant families and the evidence on the importance of home environment in early childhood, this study investigates factors related to the language and literacy resources of Turkish families in England.

Methods
As part of the European Commission-funded Horizon 2020 Project: Inclusive Education and Social Support to Tackle Inequalities in Society (ISOTIS; www.isotis.org), non-native families in 10 countries participated in a survey study investigating the strengths and potentials of culturally and linguistically diverse immigrant families from Turkish, Maghrebi and Romani backgrounds (N = 2495) (Broekhuizen, Ereky-Stevens, Wolf, & Moser, 2018). The current study focuses on the Turkish families sampled in England (N = 168). Participants were recruited by Turkish-speaking researchers using purposive and snowball sampling in community centres, supplementary schools, community events and online networks for Turkish-speaking people. Ethical approval was received from the Oxford Central University Research Ethics Committee following the Oxford University Ethical Guidelines (CUREC, 2018).

Participants
168 first-generation Turkish parents living in Greater London (N = 117) and Northwest England (N = 51) participated in this survey. The parents were the primary caregivers of the child and responsible for childcare at home. A majority of the parents were mothers (N = 143, 85%); and a small minority were fathers (N = 25, 15%). The age of parents ranged from 24 to 51 (M = 36.95, SD = 5.81).

Procedures
Interviews with parents were carried out by trained Turkish-speaking researchers at homes (41%), community centres (43%) or through phone calls (16%) depending on the preference and availability of the participants. The interviews were administered using Lime Survey either simultaneously on laptops (47%) or a paper-pencil interview (53%) and took an hour on average. During laptop interviews, the answers were entered on the Lime Survey simultaneously by the researchers. During paper-pencil interviews, the same interview procedure was followed with answers noted on a printed questionnaire and later entered into Lime Survey. Most questions required answers on a scale (e.g. 1 to 5 scale ranging from 'disagree' to 'agree'). A set of show cards was made available to parents throughout the interview to visualise the answer scales. More information on data collection procedures can be found in the technical report of the parent survey of the ISOTIS project (Broekhuizen, Ereky-Stevens, Wolf, & Moser, 2018).
The information on family background characteristics and the resources in home environment was collected through a set of questions developed by ISOTIS researchers from diverse academic backgrounds (e.g. education, psychology, sociology and economics). The questions were chosen drawing on questionnaires that had been successfully used in previous studies to assess constructs such as family socioeconomic background, migration history and home learning environment (Broekhuizen, Ereky-Stevens, Wolf, & Moser, 2018), and the questions were translated into Turkish, with back translation to ensure translation validity.

Measures
Parent's Education. To assess education level, parents received questions about their years of education, age and grade of leaving school, vocational training/ diploma and university attendance/diploma and the data were recorded nominally. Using a syntax, the information was later categorised into a fivelevel scale based on the International Standard Classification of Education (ISCED) 2011: (1) ISCED 0 and 1 (primary education or lower), (2) ISCED 2 and 3 (lower or upper secondary education), (3) ISCED 4 and 5 (postsecondary non-tertiary and short cycle tertiary education), (4) ISCED 6 (bachelor level) and (5) ISCED 7 (graduate level education).
Material Well-Being. The data on the material well-being of the families came from information on families' monthly income and material assets, which were aggregated to create a composite score of material well-being by adding the scores of monthly income and material assets of the families. To assess the material assets of families, parents received 13 questions on material resource indicators developed by the European Union (EU) (Guio et al., 2016). The indicators for material assets and resources assessed the availability of family resources for paid activities (e.g. going for a dinner once a month) or expenditures (e.g. going for a dinner once a month and replacing broken furniture), and ownership of material assets (e.g. owning a car), coded on a dichotomous level (1 = can afford and 0 = cannot afford). To assess the monthly family income, a scale developed by the ISOTIS researchers was used. The information on family income was captured using a monthly income scale with 10 options ranging from '< £900' to '> £4880 or more'. The indicators demonstrated an acceptable Cronbach's alpha coefficient of .70 and were summed to create a material well-being variable.
Social Integration. The data on the social integration of parents came from questions on parent's length of residence and English competence. To assess the length of residence, parents received questions on their country of birth and year of arrival in Britain. Lastly, to assess parent's language competence, parents were asked to indicate their language skills in English via three questions on their ease with having conversations, reading newspapers and listening to the news on a 1 to 6 interval scale ranging from 'never have problems' to 'I cannot read/speak'. Different components of parent's English skills were highly correlated with each other, whereas parent's language skills have shown an acceptable level of correlation with parent's length of residence. The data were analysed to create a composite score of social integration by assigning weighted scores produced by the principal component analysis (PCA). To ensure the appropriateness of PCA, the assumptions of the analysis was examined. The Kaiser-Meyer-Olkin measure of sampling adequacy attained a value of .756, above the recommended value of .5. Bartlett's Test of Sphericity was statistically significance (p < .001), supporting the factorability of the correlation matrix. The factor created as a final measure of parent's social integration established an eigenvalue of 2.845 and explained 71% of the variance, and the weighted scores from the analysis were used as the social integration scores of parents.
Language and Literacy Activities. The information on language and literacy activities at home was collected through questions on the home learning environment. For the purposes of this study, parent's answers to three types of activities that have been shown to be the significant predictors of children's language development have been selected (for a review, see Bus et al., 1995). These items assessed how often parents (1) had daily conversations with their children; (2) had conversations with their child on the child's general interests and (3) had shared book-reading activities with their child. For each type of activity, participants reported their frequency of participating on a 1 to 6 interval scale from 'almost never' to 'everyday'. As the items were significantly correlated with each other (r > .30), the scores for each item were summed to create the language and literacy activities variable for the final analysis.
Home Language Use. At the end of the questions on activities, parents received two questions on their language use separately for language and literacy activities. Parents reported their language use on a 1 to 5 scale from 'mainly English' to 'mainly Turkish or the other language'. Parent's answers were significantly and positively correlated with each other [r = .455, N = 168, p < .001]. Hence, the scores were summed to create the home language use variable for the final analysis.

Analysis Strategy
To investigate the extent to which the family background characteristics of Turkish families explain parents' resources in child rearing a hierarchical regression model was developed drawing on the Family Investment Model (Conger & Donnellan, 2007). Regression analyses were carried out using the parent's education, family material well-being and social integration as the predictors and the measures for language and literacy activities and home language use as the outcome variables. In these models, three blocks were developed to depict the different aspects of family background characteristics. In the first block, parent's education was taken as the independent variable of primary interest. The second block included the family material well-being, and the third block included social integration variable. For the analyses, IBM SPSS 23.0 was used.

Descriptive Statistics
See Table 1 for the descriptive statistics of the variables.
Family Background Characteristics. Almost a half of the parents had at least a bachelor's degree (N = 80, 48%). Less than a quarter of parents had a vocational degree (N = 31, 18%) or had only school qualification (N = 34, 20%). Only a small minority of the parents did not have a qualification or only had a primary school education (N = 23, 14%). For material well-being of families aggregated scores ranged from 2 to 23 with a mean score of 17.38 (SD = 4.96) and no outliers. For social integration of parents, the weighted scores ranged from À.249 to 1.44 with a mean score of .00 (SD = 1.00) with no outliers.
Language and Literacy Activities. The aggregated scores for language and literacy activities ranged from 8 to 18 with a mean score of 16.01 (SD = 2.19). The Shapiro-Wilk statistics was significant (p < .001), and the data were negatively skewed, all of which indicating a non-normal distribution of the aggregated scores. As the logarithmic transformation did not lead to a significant change in the distribution of the scores, the aggregated scores were retained for the regression analysis. To repeat the regression analysis with binomial regression to ensure the validity of the findings, an alternative variable was also developed by assigning participants into low-, medium-and high-frequency language and literacy activity households.
Home Language Use. The aggregated scores of the reported language use in conversation and literacy activities ranged from 2 to 10 with a mean score of 6.02 (SD = 2.42) with no outliers. The values for skewness and kurtosis were in the acceptable range and the data did not demonstrate outliers. Correlations between the family background characteristics and outcome variables indicated significant associations. Parent's education level was Similarly, material well-being of families was positively correlated with English use in language and literacy activities [r = À.226, N = 168, p = .003]. Parent's level of social integration was highly correlated with their language use in language and literacy activities [r = À.527, N = 168, p < .001], such that parents with higher social integration scores tended to use English more often with their children.

Bivariate Correlations
Correlations between the outcome variables indicated associations between the outcome variables, although the correlations were low, for example, the correlation between the parent's English use in language and literacy activities and their participation in the activities [r = À.141, N = 168, p = .068]. See Table 2 for the Pearson's correlation coefficients.

Regression Models
Language and Literacy Activities. A hierarchical multiple regression analysis was run to examine whether the family background characteristics predicted the language and literacy activities at home. After the entry of block 1 (parent's education), the proportion of variance explained was 17% and parent's education was shown to have a significant effect on home language and literacy activities [F(1, 166) = 33.233, p < .001]. After block 2 had been entered into the model, the model as a whole explained 17% of the variance and this model was also significant [F(2, 165) = 16.588, p < .001]. However, the R square change value was null and insignificant (ΔR 2 = .001, p = .729), which suggested that when the effect of parent's education was controlled for, material well-being did not have an additional significant effect on the learning resources of families. Lastly, when the social integration variable was entered in the block 3, the model as a whole explained 18% of the variance and this model was also significant [F(3, 164) = 11.990, p < .001]. Again, the R square change value was not significant (ΔR 2 = .012, p = .116). Due to skewness in the data, a binomial regression analysis was carried out using the three-level variable (low, medium and high) for language and literacy activities to ensure the results of the regression analysis was reliable. The final model significantly explained 11 % of the variance of the high-frequency language and literacy activities in the home environment (X 2 (3) = 13.105, p = .004), see Supplementary Appendix A for this alternative analysis. See Table 3 for the hierarchical regression outcomes. Only parent's education made a significant contribution to all three models. In the first model, parent's education was significant with a positive beta value (β = .705, p < .001). Similarly, the values were positively significant in the second (β = .682, p < .001) and third model (β = .624, p < .001) with a small decrease in its effect size. The positive beta values indicated that parents with higher education backgrounds were more likely to participate in language and literacy activities with their children. The binomial regression analysis also confirmed that parent's education was a significant and positive predictor of high-frequency language and literacy activities in the home environment (β = .499, p = .002). Home Language Use. The regression analysis was repeated with the home language variable as the outcome variable. After the entry of block 1 (parent's education), the proportion of variance explained was 3% and parent's education demonstrated a significant effect on home language use [F(1, 166) = 5.095, p = .025]. After block 2 had been entered into the model, the model as a whole explained 6% of the variance and this model was also significant [F(2, 165) = 5.443, p < .005]. The R square change value was also significant (ΔR 2 = .032, p = .019), which suggested that when the effect of parent's education was controlled for, material resources had a significant additional effect in explaining the home language. Lastly, when the social integration variable was entered in the block 3, the model as a whole explained 29% of the variance and the addition of this variable significantly improved the model predicting the home language use [F(3, 164) = 21.083, p < .001]. Moreover, the R square change value in this regression analysis was significant (ΔR 2 = .216, p < .001), suggesting that family social integration was a significant predictor of home language use after controlling for family socioeconomic background. See Table 4 below for the regression analysis results. All three of the variables in the model made a significant contribution predicting the home language use. In the first model, parent's education was .29*** ΔR 2 .23*** significant with a negative coefficient value (β = À.328, p = .025). When the variable for material resources was added in the model, the effect of parent's education predicting the home language diminished (β = À.261, p = .074), and material resources became a significant contributor (β = À.098, p = .009). When the variable for social integration was added in the model, both parent's education (β = .065, p = .636) and material resources (β = .000, p = .989) became insignificant. Social integration, then, had a significant effect in the model (β = À1.303, p < .001). The negative beta values indicating that increase in parent's education, material resources and social integration predicted increased English use at home.

Discussion
The first goal of this study was to examine the characteristics of Turkishspeaking families in England and to determine how they may affect children's home language and literacy environments. Drawing on the Family Investment Model (Conger & Donnellan, 2007), this study investigated the relative contribution of socioeconomic background on Turkish families in England. According to the model, the indicators of economic hardship such as low income, work instability and material deprivation place extra stress on the family well-being, which in turn negatively affect child-rearing practices at home. By contrast, increased levels of education and financial stability provide tangible (e.g. learning/reading materials) and intangible resources (e.g. knowledge and time) for parents to support their child's development.
The findings of this study provide supporting evidence for the Family Investment Model. As shown by the correlation analyses, parents from higher education backgrounds did better financially and demonstrated higher scores on social integration. These findings suggest that the educational capital of immigrant families can play a positive role in their social and economic integration in the country. Parent's proficiency in a country's national language is a further key factor of integration. As shown by the social integration variable used in this study, the language skills of parents strongly correlated with parent's length of residence in England. Supporting previous evidence from immigrant populations including Turkish communities, this finding suggests that the time spent in the country was an important component of their linguistic integration (Driessen et al., 2002;Willard et al., 2015;Winsler et al., 2014). Lastly, families' social integration scores demonstrated a strong and positive relationship with family material well-being. As previous studies have demonstrated, social integration is an important aspect of the socioeconomic well-being of immigrant families (e.g. Dixon et al., 2012). These findings, then, indicate that understanding the socioeconomic experiences of immigrant families requires consideration of their migration backgrounds and acculturation experiences. By including these indicators, research can more dynamically capture and account for the sociodemographic factors related to parental engagement at home.
Analyses also focused on the relative contribution of family background indicators in explaining the resources available in the home language and literacy environment. The results have shown that the parent's education background was a significant predictor of the language and literacy activities at home, whereas material well-being and social integration were not. Supporting the Family Investment Model (Conger & Donnellan, 2007), this finding has shown that parents from higher education backgrounds participated more often in learning activities that are beneficial for their child's language and literacy development. Although the social integration variable was not significant in predicting language and literacy activities, there was a significant correlation between the two. It is likely that parents' education outweighed the effect of social integration in the model due to its direct and indirect interactions with the aspects of parenting related to social integration such as English competence. Material well-being, on the other hand, was not significant in predicting parental engagement in language and literacy activities, suggesting that economic resources did not necessarily provide an advantage for the learning experiences of Turkish immigrant children. By contrast, other social and educational aspects of family background such as parents' education background did play a role. While the importance of parent's socioeconomic background in early childhood is well-known in existing literature (e.g. Conger & Donnellan, 2007), the role of social integration is still largely under-researched. In the context of migration, it is particularly important to consider the role of social integration for parenting behaviours. As Becker (2011) has shown with Turkish immigrant parents in Germany, social integration can act as a 'scaffold' for ethnic minority children's upbringing, easing their access to national language education, and parental education background can buffer the effects of migration by providing parents the knowledge and skills in supporting their child's development and learning at school.
The second goal of this study was to address how family background characteristics influenced parents' home language preferences. Even without taking the social integration variable into account, socioeconomic factors were significant in predicting at-home language use. Families with limited financial and educational resources were more likely to use Turkish as the home language which corresponds with studies on the language behaviours of Turkish parents in other European countries (Prevoo et al., 2014;Willard et al., 2015;Winsler et al., 2014). However, it is important to note that after adding the social integration variable to the model, the effects of parental education and material well-being became insignificant. In the final model, social integration was a significant predictor of language use, suggesting that parents' cultural experiences in the UK played a more profound role in determining the language environment at home. Similar findings have been captured in other studies on immigrant families (e.g. Winsler et al., 2014;Willard et al., 2015). A study on Turkish children in Germany has shown that mother's generational status in the country was a stronger predictor of the home language than mother's education level (Willard et al., 2015). While this finding might look promising for the heritage language maintenance of minority children, it is important to consider that the degree of a given family's social integration will play a significant role in parents' capacities to support their children's English language development when they start school. This means that children growing up in immigrant households with low social integration may have lower parental support in their education experiences. This may, furthermore, lead to an increase in the discrepancy between school and home language and negatively affect pupils' school attainment in the later stages of education. In fact, this has been shown for children growing up in first-generation immigrant households in the United States (e.g. Winsler et al., 2014).
To address the extent to which family background characteristics explain parent's child-rearing resources, the variance explained by the model was examined. The results of the final model demonstrated that the family background characteristics explained the highest amount of variance for language use (29%), which was followed by language and literacy activities at home (18%). A high proportion of variance explained for language use was expected as almost all studies on immigrant families have shown that parent language behaviours were predicted strongly by the indicators of socioeconomic status and migration history (Westeren et al., 2018;Willard et al. 2015;Winsler et al., 2014). The Turkish population in England, however, has not been previously studied on this issue. This study is the first to contribute to demonstrate how the socioeconomic and demographic characteristics of Turkish families in England explain their language use at home. The model was not as strong in explaining the variance in language and literacy activities. Possible reasons for this might be, first, the multifaceted nature of learning activities that take place in culturally diverse households, and second, increased awareness on child development across the minority families in England. In fact, fieldwork within Turkish community centres and private homes revealed that Turkish parents were highly invested in their children's education regardless of their socioeconomic background (ISOTIS fieldwork notes, 2018). The majority of the parents were observed to spend their weekends at community centres, often recruiting tutors and finding extracurricular opportunities to support their children's learning (see also Lytra, 2015). As this study suggests, support within the community paired with parents' observed concern for their child's academic future, might have offset the effects of socioeconomic disadvantage on at-home language and literacy development.
The findings of this study have shown that family background characteristics were important predictors of the resources in children's home language and literacy environments. However, the strength and significance of the model varied for each aspect of the home environment. Nevertheless, the evidence from this study shows that low education levels and social integration might have a negative effect on parental engagement in at-home language and literacy resources. These may be exacerbated by an increased demand in English competence as children become more proficient in the language.
These findings underline that socioeconomic disadvantages experienced by first-generation immigrant parents, especially when they are new to a country, can lead to diminished stimulation in their home environment. Challenges that migration bring into the family life can have negative effects on the family resources in child-rearing. These effects may be exacerbated if parents are new to the country or experience problems with social integration, which might result from language barriers and lack of educational resources. These issues may deprive parents from participating in their child's learning through essential learning activities and parenting practices in the early years of life.

Limitations
This study had some limitations. First, there were some challenges specific to the data collection methods that have been experienced by the researchers. One issue with administering questionnaires to parents was the non-response rate to questions on sensitive information such as monthly income. In order to minimise this limitation, the majority of the interviews were carried out faceto-face by Turkish mothers from the community. However, this issue did not completely solve problems such as false response and, in fact, elucidated it. The researchers realised that some participants provided incorrect information on questions regarding their marital status or family income as they were sceptical about the consequences of enclosing information that might affect their social benefits (ISOTIS fieldwork notes, 2018). Another issue that arose with the questionnaire use was the social desirability bias, which might have been exacerbated as a result of the participants' minority background (e.g. Swain, Heyman, & Gillman, 1998). In this study, the majority of the parents reported to participate in language and literacy activities more often than was observed for parents in the previous use of the questionnaire (Melhuish et al., 2001(Melhuish et al., , 2008. The study had some analytical limitations as well. First, a considerable amount of the variance remained unexplained for the language and literacy activities. More information on family background characteristics and their lives in England could have further elucidated the factors relevant to parental engagement in the language and literacy activities at home. For instance, including factors such as parental well-being and self-efficacy might have improved the model (e.g. Baydar et al., 2014). Second, due to limited resources in data collection, the data analysis mostly focused on the information on primary caregivers, limiting the understanding of the family background characteristics and home environment from the perspectives of the other family members. Including information on both mothers and fathers could have contributed to the analytical strength of the model in predicting the resources and languages provided for children at home and understanding the beneficial parenting practices for both mothers and fathers. Third, the families were relatively advantaged in terms of educational qualifications and material circumstances. Different findings might be found in a more disadvantaged sample.

Future Directions
The findings of this study suggest that family background characteristics account for a substantial amount of variation in the language and literacy environment of immigrant families. Being new to a country and facing linguistic barriers can limit parents' language use and access to learning resources at home. This finding is crucial for policymakers and professionals working on the integration of immigrant families. One approach to support immigrant children, therefore, should include providing newly arrived families with access to training programmes that support their children's language and literacy environment. Providing English language training to parents can facilitate their integration and have positive effects on parental engagement at home. Parents can also benefit from learning about the value of child-directed speech and language enriching activities at home in participating in their child's bilingual language learning. Another approach should therefore be to improve the inclusive strategies schools take to support the English language learning of children from first-generation immigrant families. Schools and education centres can also guide immigrant parents in their use of heritage language literacy resources at home. These could help them acknowledge and value their cultural heritage and support the bilingual language acquisition of their children.

Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This project has received funding from the European Union' s Horizon 2020 research and innovation programme under the grant agreement No. 727069.

Supplemental Material
Supplemental material for this article is available online.