Preschoolers' English vocabulary development: The influence of language proficiency and at-risk factors

Learning and Individual Differences 35 (2014) 79–86

Contents lists available at ScienceDirect

Learning and Individual Differences journal homepage: www.elsevier.com/locate/lindif

Preschoolers' English vocabulary development: The influence of language proficiency and at-risk factors Jeannette Mancilla-Martinez a,⁎, Joanna A. Christodoulou b, Michelle M. Shabaker c a b c

University of California, Irvine, United States MGH Institute of Health Professions, United States University of Illinois, Chicago, United States

a r t i c l e

i n f o

Article history: Received 2 April 2013 Received in revised form 12 April 2014 Accepted 28 June 2014 Keywords: English language learners Vocabulary development Risk factors Preschool Individual Growth Modeling

a b s t r a c t A persistent achievement gap, particularly in language development, exists between English language learners and their peers prior to kindergarten entry (Fuller et al., 2009) and throughout the formal education years (García & Frede, 2010). While language experience is known to contribute to this gap, the impact of additional risk factors, whether indexed cumulatively or individually, is not well understood. This longitudinal study investigates preschool children's (n = 204) patterns of English receptive vocabulary development by the level of English language proficiency designation at preschool entry, as well as by the influence of cumulative and individual risk factors drawn from 29 child, parent, and context variables. Our results show that, although there was positive vocabulary growth for all preschoolers, there was a more adverse impact of cumulative and individual risk for children designated as less versus more English proficient. Implications for policy, practice, and further research are discussed. Published by Elsevier Inc.

1. Introduction The achievement gap is well documented between children from non-native English speaking homes, known as English language learners (ELLs), and children from low-income homes, as compared to their peers who are monolingual and from middle- and upper-income homes, respectively (e.g., August & Shanahan, 2006). For ELLs, learning two or more languages concurrently is not solely a risk factor for academic difficulties (De Houwer, 1999; Snow, 1992). However, low socio-economic status hinders child development in academic, neurocognitive, socio-emotional, and physical health domains (Brooks-Gunn & Duncan, 1997; Conger & Donnellan, 2007; Farah et al., 2006), with the greatest impact on children in early childhood (Duncan, Ziol-Guest, & Kalil, 2010; Hernandez, 2004). This is of grave concern as Latino children from Spanish-speaking homes—the largest and fastest growing segment of the U.S. population (Passel, Cohn, & Lopez, 2011)—are now the largest single group of poor children in the U.S. (López & Velasco, 2011). A growing body of empirical research shows that the achievement gap, particularly in vocabulary development, between Spanish-speaking ELLs from low-income homes and their peers is already evident prior to kindergarten entry (Fuller et al., 2009; Mancilla-Martinez & Lesaux, 2011a) and continues at every level of the education system (August & Shanahan, 2006; García & ⁎ Corresponding author at: University of California, Irvine, School of Education, Irvine, CA 92617, United States. Tel.: +1949 824 2672 (office). E-mail address: [email protected] (J. Mancilla-Martinez).

http://dx.doi.org/10.1016/j.lindif.2014.06.008 1041-6080/Published by Elsevier Inc.

Frede, 2010). In light of the well-established link between vocabulary and overall academic achievement (e.g., Anderson & Freebody, 1983; Snow & Kim, 2007), attending to the vocabulary needs of the young Latino population—particularly those from low-income homes—is urgent (García & Frede, 2010; García & Jensen, 2009; National Task Force on Early Childhood Education for Hispanics, 2007). Despite the consistent finding that poverty is associated with the low academic achievement of many ELLs, we know very little about the influence of other factors on the vocabulary gap during the preschool years, whether these factors are considered alone or in combination, they are directly relevant to the child, or they are indirectly relevant to the child via parents or context. Previous research has used cumulative risk models to demonstrate that vulnerable populations are disproportionately burdened by multiple risk factors (Burchinal, Vernon-Feagans, & Cox, 2008; Cadima, McWilliam, & Leal, 2010; Garmezy, Masten, & Tellegen, 1984; Rutter, 1979). Though cumulative risk research primarily has been conducted with monolingual populations, some work shows that children from immigrant families are more than twice as likely to experience multiple risk factors than those from native-born families (Hernandez, 2004). While it seems intuitive that English language proficiency contributes to the noted vocabulary achievement gap between ELLs and their non-ELL peers, to our knowledge, the role of cumulative risk for vocabulary development in preschool-age ELLs has yet to be studied. Our longitudinal study investigated the English vocabulary development of children enrolled in an English-only state-funded preschool program over the course of the academic year. The influence of children's

80

J. Mancilla-Martinez et al. / Learning and Individual Differences 35 (2014) 79–86

English language proficiency status and risk factors at preschool entry on vocabulary was explored. By following the vocabulary development of children from both monolingual English and Spanish–English bilingual homes, this study provides insight into factors that shape the English vocabulary development of young children, above and beyond the role of initial English language proficiency designation, which is typically the dominant focus of research with respect to ELLs. 1.1. Language proficiency designation and vocabulary in ELLs English language proficiency assessments are used to ensure that schools and districts are held accountable for ELL progress toward achieving English language proficiency. The assessments are used to identify language proficiency designations, thereby assisting schools in the identification of appropriate classroom placements for students. Language proficiency encompasses diverse components, such as listening comprehension, grammatical skills, vocabulary, and oral retelling skills. The IDEA Oral Language Proficiency Tests (IPT), available for students in preschool through grade 12, is an example of a measure used to determine students' overall English language proficiency status. The IPT includes items across language constructs that are designed for screening purposes. Therefore, a student's language proficiency designation represents a general proxy for language proficiency. In contrast, measures of vocabulary, for example, represent proficiency within a specific language component. Differentiating between English language proficiency and vocabulary knowledge is of central importance in considering children from non-native English speaking homes, some of whom are considered ELLs, as mounting evidence points to their generally depressed vocabulary levels beginning in early childhood and beyond (August & Shanahan, 2006; Fuller et al., 2009; García & Frede, 2010; Mancilla-Martinez & Lesaux, 2011a). 1.2. Risk factors While investigations to date have examined the impact of risk factors, cumulatively, in selected populations, the potential influence of cumulative risk factors for ELLs remains underspecified. Cumulative risk models, considered to be important for understanding how children respond to stressful factors in their lives (Rutter, 1979), have been examined in low-income European–American infants (Burchinal et al., 2008), Portuguese Caucasian preschoolers (Cadima et al., 2010), and children from one of the three stressful contexts (urban settings; having had a stressful early health defect; children with physical disabilities transitioning to mainstream schooling) (Garmezy et al., 1984). Further evidence supporting cumulative risk investigations comes from the U.S. Centers for Disease Control and Prevention ACE study, which points to the lifelong impact of adverse childhood experiences on health and social outcomes, with graded increases associated with more risk factors (Felitti et al., 1998). Evidence suggests that children from low-income homes show a cumulative negative impact on preschool language skills based on risk factors drawn from maternal, paternal, and birth data (Stanton-Chapman, Chapman, Kaiser, & Hancock, 2004). Thus, while studies have used cumulative risk models in the past, an integrative evaluation of ELLs remains a critically underinvestigated area. Furthermore, employing longitudinal designs to study the interaction between a child's development and relevant contextual factors offers the ability to document significant factors and the interactions that influence risk status. Currently, few studies have explored the impact of multiple risk factors, but evidence shows that lower income status is associated with more co-occurring risk factors (Evans & Kim, 2010). The degree to which ELLs may be disadvantaged for vocabulary development may be exacerbated by cumulative risks related to income status and more broadly by child, parent, and context risk factors; it is critical to study their impacts in this understudied population early in development.

To better understand how the amount and type of risk affect outcomes, the contributions of cumulative and individual risk factors must be evaluated (Burchinal, Roberts, Hooper, & Zeisel, 2000). Such work is sorely needed in relation to ELLs and their vocabulary development if we are to gain a more nuanced understanding of factors that influence their vocabulary development above and beyond English language proficiency. The purpose of this longitudinal study, conducted in English-only classrooms, was to investigate preschool children's vocabulary development over the course of the preschool year and study the impact of and possible interactions with additional risk factors. Potential differential growth patterns due to children's English language proficiency designation at preschool entry and the extent of applicable risk factors drawn from child, parent, and context levels were explored. We asked: 1) What are the patterns of English vocabulary development among preschool children from low-income homes, compared to national monolingual norms, and to what extent do they vary by English language proficiency designation at preschool entry? 2) Accounting for an initial level of English language proficiency, what is the influence of cumulative risk (i.e., total number of risk factors) on English vocabulary initial levels and rates of growth? 3) Accounting for an initial level of English language proficiency, which individual risk factors (e.g., low parental education level, multiple families in the home) influence English vocabulary initial levels and rates of growth?

2. Method 2.1. Study design All children (n = 204) enrolled in English-only classrooms at a public, half-day preschool program in Illinois during the 2011–2012 academic year participated in this study. Participating children were followed for the duration of their preschool year. There was minimal attrition from the cohort over time. Of the 204 students who participated in the fall, all were assessed in the winter (0% attrition) while 14 were not assessed in spring (n = 190; 7% attrition). 2.2. Participants Demographic data was provided by the district. All children were born in the U.S. (n = 204) whereas 126 (62%) mothers were born in the U.S. mainland, 70 (34%) were born in Mexico, and eight (4%) were born outside of the U.S. mainland and in other countries. Of the fathers with country of birth information available (n = 162), 74 (46%) were born in the U.S. mainland, 83 (51%) were born in Mexico, and five (3%) were born outside of the U.S. mainland and in other countries. The primary home language was Spanish for 126 families (62%) and English for 78 families (38%). Of the families who completed an application for free or reduced lunch (n = 166), 90% qualified (78% for free and 11% for reduced lunch). Thus, as a group, the children are all U.S.-born from predominantly low-income households. 2.3. Procedure At-risk status was determined via the preschool screening process, English language proficiency designation was determined via direct assessment at preschool entry, and receptive vocabulary was assessed one-on-one at three time points: fall, winter, and spring of the preschool year (see Table 1 for testing ages). Lead teachers (n = 13), trained by the first author on the administration of the vocabulary assessment during 2 separate, 3-hour training sessions, administered the assessments in a quiet room during each 2-week testing period.

J. Mancilla-Martinez et al. / Learning and Individual Differences 35 (2014) 79–86 Table 1 Average age of testing at each measurement point, with standard deviation in parentheses.

Fall of preschool Winter of preschool Spring of preschool

N

Age in months

204 204 190

50.6 (6.8) 54.3 (6.9) 57.7 (6.8)

Note. Age in months is calculated based on the birth date in relation to the test date.

2.4. Measures 2.4.1. At-risk classification: cumulative index and individual Prior to preschool entry, all children and families were required to complete a screening to determine eligibility to attend the program. Families participated in a one-on-one eligibility interview with a member of the preschool staff. The interviewer asked each family a series of questions regarding 29 at-risk factors (see Table 2) in English or Spanish, per parental preference, and the applicability of these factors to the child/family. Each factor was dichotomously scored (0 = does not apply, 1 = applies). We generated a Cumulative Risk Index, which entails the simple sum of dichotomized risk factors and is common in studies focused on child outcomes (e.g., Burchinal et al., 2000; NICHD Early Child Care Research Network, 2004). The full sample average was 8 risk factors (SD = 2). Additionally, we focused on individual risk factors that applied to 10% or more of the full sample as some of the risk factors applied to less than 1% of the sample (e.g., lead poisoning) and would not account for variation in children's vocabulary development. This yielded 20 individual at-risk factors for inclusion (see factors set in bold, Table 2).

81

2.4.2. English language proficiency designation The district individually administers the Pre-IPT (Idea Proficiency Test)–Oral English Test (Ballard, Tighe, & Dalton, 1979) to all students as the screener for initial designation as Non-English (NEP), Limited English (LEP), or Fluent English Proficient (FEP). The Pre-IPT assesses proficiency in four domains of oral language: vocabulary, grammar, comprehension, and verbal expression. The items are grouped into levels, and the examiner scores the student's answers as correct or incorrect. The students advance through the test levels until the test is completed or until they stop at the highest attainable level based on their language proficiency. Again, the children receive a specific designation rather than a continuous score. Reliability estimates at the subtest level range from .85 to .91. Children from EO homes (n = 78; 38% of the sample) did not complete the Pre-IPT and were automatically classified as EO speakers. Of the children from Spanish-speaking homes, 46 were identified as NEP (23% of the sample), 66 as LEP (32% of the sample), and 14 as FEP (7% of the sample). Thus, a total of 4 language groups were identified. 2.4.3. Receptive vocabulary The Peabody Picture Vocabulary Test (PPVT-IV; Dunn & Dunn, 2007) was used to assess children's English receptive vocabulary during fall, winter, and spring. Children were required to point to the picture that matched the target word provided by the examiner. The test–retest reliability is .92. 2.5. Analytic approach To examine patterns of development in receptive vocabulary across the four language groups, we used Individual Growth Modeling using

Table 2 Descriptions of at-risk factors by category, with percentage of students for whom the risk factor is applicable (n = 204). At-risk factors

Descriptions

Full NEP LEP FEP EO (n = 204) (n = 46) (n = 66) (n = 14) (n = 78)

Active IEP Non-biological At-risk referral Child health

Does the child have an active Individual Education Plan Is this a foster or adopted child? Has the child been identified as at-risk for developmental delays? Is there a history of or a medical concern (e.g., asthma, lack of immunizations)? Did the child participate in any early intervention program? Has the child had lead poisoning? Was the child born prematurely? Are the special circumstances such as living in different places that affect the child? Has the child been identified as at-risk for special education placement? Is a parent on active military duty? Was a parent under age 20 at the time of the child's birth? Is a parent receiving counseling services? Does a parent drink excessively and/or use illegal drugs? Has a parent been involved with the justice system? Is a parent limited English proficient? Does a parent have less than a high school education? Does a parent have an identified learning difficulty (e.g., dyslexia)? Is someone other than a parent the primary caregiver? Is the child being raised by a parent single? Is a parent temporarily unemployed or unemployed? Are there more than 3 children under age 5 living in the household? Does the family receive agency support (e.g., WIC)? Is there a family history of or medical concern (e.g., heart disease, diabetes)? Does the family qualify as low-income? Has the family moved more than 2 times during the last 4 years? Are multiple families living in the same household? Does a child's sibling have an Individualized Education Plan? Did a child's sibling attend the same preschool program? Does the family live in subsidized housing?

8 b1 63 19

7 0 53 13

6 2 80 12

0 0 43 14

13 1 59 28

10 b1 18 16

11 0 13 20

15 0 18 14

14 7 7 14

4 0 23 15

15

36

6

14

12

b1 56 5 b1 19 52 56 11 2 56 49 8

0 47 2 2 7 89 82 2 0 38 51 7

0 58 6 0 18 76 58 9 2 52 44 9

0 50 7 0 21 50 50 7 0 43 21 0

1 62 5 1 26 12 41 17 4 62 55 9

85 19

84 22

88 21

79 14

83 15

89 22 40 15 36 5

93 20 33 24 49 0

89 23 45 11 27 0

79 14 36 7 57 0

87 26 38 15 31 12

8

8

8

7

8

Early intervention Lead poisoning Premature Special circumstances Special education referral Parent military Adolescent parent Counseling services Substance abuse Criminal involvement Limited English Education level Learning difficulty Primary caregiver Single parent Employment Young children Agency support Family health

Low income Mobility Multiple families Sibling IEP Sibling preschool Subsidized housing ————————————————————————— Cumulative Risk Index Average number of risk factors

82

J. Mancilla-Martinez et al. / Learning and Individual Differences 35 (2014) 79–86

the multilevel model for change (Singer & Willett, 2003), with age in months used to index time. The analyses were conducted in a person– period dataset that contained the longitudinal data on all sampled children, using SAS PROC MIXED with full maximum likelihood estimation. Per Singer and Willett, the likelihood ratio test was used as the primary criterion for evaluating model fit. For the outcome variable measuring growth, we used the Growth Scale Value (GSV), a developmental scale score for the PPVT-IV that has been vertically equated using Item Response Theory. The GSV indexes absolute growth rather than growth relative to the norm, essential for studying inter-individual differences in change over time. The GSV is scaled such that a score of 120 corresponds to the performance of an average 4-year-old. Receptive vocabulary was measured at three time points, limiting us to a linear functional form. Inspection of empirical growth plots of each child's vocabulary scores as a function of age indeed suggested linear growth, as suggested based on students' mean scores from fall to spring (see Table 3) and convergent with studies of young children's developing vocabularies (e.g. Hart & Risley, 1995; Mancilla-Martinez & Vagh, 2013; Pan, Rowe, Singer, & Snow, 2005). Thus, for instance, the following model was fit: EVocabij ¼ γ00 þ γ10 ðChildAge−50Þij þ γ01 NEP þ γ11 NEP ðChildAge−50Þij þ γ02 LEP þ γ12 LEP ðChildAge−50Þij þ γ03 FEP þ γ13 FEP ðChildAge−50Þij þ h i ζ 01 þ ζ 00 ðChildAge−50Þ þ ε ij

#!     " 2   σ 0 σ 01 0 ζ where ε ij  N 0; σ 2ε and 0i  N ; . 0 ζ 1i σ 10 σ 21 The subtraction of 50 from child age allowed for a meaningful interpretation of the parameter estimates: γ00 represented the average score for children at the first measurement point (i.e., preschool entry) and γ10 represented the average true initial, instantaneous slope. Parameters γ01, γ02 and γ03 represented the effects on initial achievement of the NEP, LEP, and the FEP groups, respectively. Parameters γ11, γ12 and γ13 represented the effects on initial rate of growth of the NEP, LEP, and the FEP groups, respectively. The random effect εij is a level-1 residual for child i at time j and is assumed to be drawn from a normal distribution with mean of 0 and unknown variance σε2. Random effects ζ0i and ζ1i represent child level-2 residuals for the intercept and slope, respectively. Table 3 Sample means on English receptive vocabulary by wave for each language group, and for the full sample, with sample standard deviations in parentheses. Language groups

N

GSV score

Standard score

English Only Fall of preschool Winter of preschool Spring of preschool

78 78 71

110.1 (18.1) 122.9 (16.4) 130.8 (15.7)

92.5 (13.5) 100.3 (12.6) 105.2 (12.6)

Fluent English Proficient Fall of preschool Winter of preschool Spring of preschool

14 14 12

113.2 (10.9) 125.0 (14.6) 129.8 (12.2)

96.3 (14.4) 103.7 (18.5) 105.0 (11.3)

Limited English Proficient Fall of preschool Winter of preschool Spring of preschool

66 66 64

106.6 (14.0) 120.2 (14.1) 129.0 (12.9)

88.9 (11.5) 98.2 (11.7) 103.0 (10.7)

Non-English Proficient Fall of preschool Winter of preschool Spring of preschool

46 46 43

84.4 (25.3) 100.3 (24.0) 109.9 (20.6)

73.0 (19.7) 83.2 (19.9) 87.3 (17.3)

Full sample average Fall of preschool Winter of preschool Spring of preschool

204 204 190

103.4 (21.1) 117.1 (19.8) 125.4 (17.9)

87.2 (16.6) 96.0 (16.2) 100.5 (15.0)

They are both hypothesized to be drawn from a multivariate normal distribution with a mean of zero, unknown variances σ20 and σ21, and unknown covariance σ01. We then investigated whether at-risk status, cumulatively and then individually, influenced students' initial vocabulary levels and/or rates of vocabulary growth over the course of the year. 3. Results 3.1. Preliminary descriptive analyses Table 3 displays students' English vocabulary GSV and standard scores for each of the groups, as well as for the full sample, across all time points. Vocabulary skills were within the average range for the EO, FEP, and LEP groups, but were nearly 2 SD below national norms in fall of preschool and in the low-average range by spring for the NEP group. For the full sample, the average fall of preschool score fell nearly 1 standard deviation below national norms, but reached the national average by spring. We also examined the distribution of at-risk factors for each of the language groups. Except for the FEP group (average of 7 risk factors), there was no difference in the average number of risk factors (8) for the other three groups and no difference in the standard deviation for any of the groups. Finally, we examined the correlation between vocabulary scores at each time point and both English language proficiency designation and cumulative risk. Students' vocabulary scores had a low-moderate, positive correlation with English language proficiency (r = .40 for fall, .37 for winter, and .36 for spring), revealing that English language proficient and receptive vocabulary knowledge were only somewhat related. Convergent with previous work examining cumulative risk and receptive language use among preschool-age children from low-income homes (Burchinal et al., 2000), there was no correlation between the vocabulary scores and cumulative risk. 3.2. Growth modeling results Given our substantive interest in potential differences in the growth trajectories of children who entered preschool with varying levels of English language proficiency, we examined the growth trajectories for the four language groups: English-Only (EO), Fluent English Proficient (FEP), Limited English Proficiency (LEP), and Non-English Proficiency (NEP). The EO group was the reference group for this analysis. To quantify the absolute magnitude of the observed differences (i.e., gaps) in children's vocabulary performance across the different language groups using a standardized metric, we calculated effect sizes at all time points by dividing the mean difference by the national norms standard deviation, allowing us to determine how many standard deviations the means of the different language groups were apart from the national norming sample, as well as from one another. This allowed us to interpret differences using Cohen's (1992) conventions for effect sizes (i.e., ~.2 a small effect, ~.5 a medium effect, and .8+ a large effect). Table 4 presents the results of a series of multilevel models fitted to represent students' English receptive vocabulary growth. The unconditional growth model (Model EV1) indicated that the linear specification of vocabulary as a function of age described the shape of vocabulary over time. The significant linear term indicates that, on average, the rate of change in students' vocabulary skills is positive. We then assessed the impact of initial language proficiency on English vocabulary initial status (fall of preschool) and rate of growth (see Table 4, Model EV2). The EO group was the reference group and the coefficients for the effects of the NEP, LEP, and FEP groups can thus be interpreted as the difference between these groups and EO children. Next, we assessed the influence of students' at-risk status, first cumulatively (see Table 4, Model EV3) and, in a separate analysis, individually (see Table 5).

J. Mancilla-Martinez et al. / Learning and Individual Differences 35 (2014) 79–86

83

Table 4 Taxonomy of growth multilevel models for change for English receptive vocabulary as a Function of Linear Age, Initial Language Proficiency and Cumulative Risk Index (n = 204). Model EV1 (unconditional growth) Fixed effects Initial status, π0i Intercept

γ00

Non-English Proficient (NEP)

γ 01

Limited English Proficient (LEP)

γ02

Fluent English Proficient (FEP)

γ 03

Cumulative Risk Index

γ 04

Non-English Proficient by Cumulative Risk Interaction

γ 05

Initial rate of change, π1i Intercept

γ 10

Non-English Proficient (NEP)

γ 11

Limited English Proficient (LEP)

γ 12

Fluent English Proficient (FEP)

γ 13

Variance components Level 1: Within-person Level 2: Between-person

σ2ε σ20

In rate of change

σ21

Covariance

σ01

Goodness of fit statistics Deviance (-2LL) AIC BIC

105.08*** (1.26)

2.61*** (0.10)

55.14*** (5.00) 283.37*** (33.81) 0.57*** (0.23) −7.12*** (1.97) 4884.5 4896.5 4916.9

Likelihood ratio tests H0 γ 01 = γ 02 = γ 03 = γ 04 = γ 05 = γ 11 = γ 12 = γ 13

Model EV2 (incl. all language groups)

Model EV3 (incl. cumulative risk)

Model EV4 (incl. interaction term)

109.44*** (1.77) −16.18*** (2.97) −2.55 (2.74) 5.37 (4.88)

120.84*** (3.70) −12.86*** (2.52) —

117.84*** (3.93) 9.88 (10.90) —

—

—

−1.67*** (0.46)

−1.28** (0.49) −2.80* (1.31)

2.54*** (0.10) —

2.54*** (0.10) —

—

—

—

—

2.42*** (0.15) 0.36 (0.25) 0.15 (0.23) −0.85* (0.40) 59.64*** (5.42) 224.27*** (28.16) 0.29~ (0.21) −4.70*** (1.67)

58.27*** (5.33) 225.28*** (29.23) 0.38* (0.22) −5.44*** (1.77)

57.68*** (5.28) 221.13*** (29.0) 0.41* (0.22) −5.52*** (1.77)

4849.9 4873.9 4914.6

4656.3 4672.3 4698.8

4651.7 4669.7 4699.6

34.6***

193.6***

4.6***

*p b .05, **p b .01, ***p b .001.

3.2.1. Research Question 1: vocabulary growth by English language proficiency designation Compared with the children classified as EO at preschool entry, LEP and FEP groups did not exhibit significantly lower initial (i.e., fall of preschool) English receptive vocabulary levels. In contrast, the NEP group had significantly lower initial vocabulary levels compared to EO chil^ 01 = −16.18, p b .001; see Table 4, Model EV2). For the rates dren (γ of vocabulary growth, the FEP group exhibited a slightly lower rate of ^ 13 = − 0.85, p b .05; see Table 4, growth compared to EO children (γ Model EV2). 3.2.2. Research Question 2: vocabulary growth by English language proficiency designation and cumulative risk Children with more risk factors had lower vocabulary levels com^ 04 = −1.67, p b .001; see pared to those with fewer risk factors ( γ Table 4, Model EV3). Further, the NEP group continued to exhibit lower initial vocabulary levels, but, accounting for cumulative risk, children from the FEP group no longer displayed lower rates of vocabulary growth. For parsimony, Table 4 only shows the significant effect of NEP on initial vocabulary levels; it does not show the non-significant effects of a) LEP and FEP on initial vocabulary levels or b) language designation on vocabulary rates of growth. We then investigated whether the influence of cumulative risk varied by students' language proficiency designation (i.e., for the NEP group compared to the other groups) and there was a ^ 05 = −2.80, p b .001; see Table 4, Model EV4). significant interaction (γ

In summary, only the NEP group displayed a lower initial level of vocabulary compared to the EO group. Additionally, there was no difference between any of the 4 language groups in the rate of vocabulary growth once cumulative risk was accounted for. Finally, there was a significant interaction between English language proficiency designation and cumulative risk. Because the LEP and FEP groups did not differ from the EO group, Fig. 1 displays the fitted vocabulary growth trajectories for the LEP, FEP, and EO groups combined (referred to as ‘English Proficient’ and depicted with the square symbol) compared to the NEP group (referred to as ‘Non-English Proficient’ and depicted with the circle symbol) at above average (i.e., 10 or more risk factors), average (i.e., 6–9 risk factors), and below average (i.e., fewer than 6 risk factors) levels of cumulative risk. For comparison, the national average is also shown (depicted by the diamond symbol). As illustrated, the English Proficient students outperform the Non-English Proficient students independent of their levels of cumulative risk. However, illustrating the effect of the significant interaction, the vocabulary gaps between English Proficient and Non-English Proficient students increase by levels of cumulative risk. For instance, in fall of preschool, the gap between English Proficient and Non-English Proficient students with below average risk is 0.63 SD, increasing to 0.67 by spring of preschool whereas the gap between English Proficient and Non-English Proficient students with above average risk is 1.14 SD in fall of preschool, increasing to 1.27 by spring of preschool.

84

J. Mancilla-Martinez et al. / Learning and Individual Differences 35 (2014) 79–86

Table 5 Taxonomy of growth multilevel models for change for English receptive vocabulary as a function of linear age, initial language proficiency and individual at-risk factors (n = 204). Model EV1 Model EV2 Model EV3 Model EV4 (unconditional growth) (incl. all language groups) (incl. individual risk factors) (incl. interaction term) Fixed effects Initial status, π0i Intercept

γ00

Non-English Proficient (NEP)

γ 01

Limited English Proficient (LEP)

γ02

Fluent English Proficient (FEP)

γ 03

Single Parent

γ 04

Multiple Families

γ 05

Parent LEP

γ 06

105.08*** (1.26)

109.44*** (1.77) −16.18*** (2.97) −2.55 (2.74) 5.37 (4.88)

112.14*** (2.25) −11.43*** (2.68) —

111.45*** (2.22) 3.77 (6.54) —

—

—

−5.73* (2.21) 5.01* (2.14) −6.47*** (2.26)

−5.63** (2.16) 4.81* (2.09) −4.64~ (2.32) −18.21* (7.14)

2.52*** (0.10) —

2.50*** (0.10) —

—

—

—

—

Non-English Proficient by Cumulative Risk Interaction Initial rate of change, π1i Intercept

γ 10

Non-English Proficient (NEP)

γ 11

Limited English Proficient (LEP)

γ 12

Fluent English Proficient (FEP)

γ 13

Variance components Level 1: Within-person

σε 2

Level 2: Between-person

σ20

In rate of change

σ21

Covariance

σ01

Goodness of fit statistics Deviance (-2LL) AIC BIC

2.61*** (0.10)

55.14*** (5.00) 283.37*** (33.81) 0.57*** (0.23) −7.12*** (1.97) 4884.5 4896.5 4916.9

Likelihood ratio tests H0 γ 01 = γ 02 = γ 03 = γ 04 = γ 05 = γ 11 = γ 12 = γ 13

2.42*** (0.15) 0.36 (0.25) 0.15 (0.23) −0.85* (0.40) 59.64*** (5.42) 224.27*** (28.16) 0.29~ (0.21) −4.70*** (1.67)

58.02*** (5.33) 225.77*** (29.39) 0.44* (0.22) −6.11*** (1.83)

58.65*** (5.38) 217.63*** (28.63) 0.42* (0.22) −6.06*** (1.77)

4849.9 4873.9 4914.6

4655.6 4675.6 4708.8

4649.3 4671.3 4707.8

34.6***

194.3***

6.3***

*p b .05, **p b .01, ***p b .001.

3.2.3. Research Question 3: vocabulary growth by English language proficiency designation and individual risk factors Three at-risk factors significantly predicted children's vocabulary growth: having a single parent, having multiple families live in the home, and having a limited English proficient parent. Children who lived in a single parent home and children whose parents were limited ^03 = −5.73, p b .05 and English proficient had lower vocabulary levels (γ ^05 = −6.47, p b .001, respectively) (see Table 5, Model EV3). However, γ children who lived in multiple-family homes evidenced higher vocabu^ 04 = 5.01, p b .05). We investigated whether the influence lary levels (γ of each of the three factors varied by the students' English language proficiency designation (English Proficient vs. Non-English Proficient) and only the interaction with parental status as limited English speaking ^ 06 = − 18.21, p b .05; see Model EV4). That is, the was significant (γ English Proficient group's vocabulary gap was .28 SD depending on parental English proficiency while the Non-English Proficient group's gap was 1.10 SD depending on parental English proficiency. Fig. 1. Interaction between initial English language proficiency designation and cumulative risk for English receptive vocabulary growth from fall to spring of preschool for students in the English-Only, Fluent English Proficient, and Limited English Proficient groups (referred to as ‘English Proficient’) and students less proficient in English (referred to as ‘Non-English Proficient’) by above average (1 standard deviation above the mean), average, and below average (1 standard deviation below the mean) cumulative risk.

4. Discussion This longitudinal study reveals that risk factors drawn from child, parent, and context levels, whether considered cumulative or individually,

J. Mancilla-Martinez et al. / Learning and Individual Differences 35 (2014) 79–86

impacted preschoolers' vocabularies beyond the initial English language proficiency designation. Of concern is that there was a more adverse impact of the risk factors investigated for the children with less English proficiency compared to their more proficient English-speaking peers, underscoring the hurdles that children who enter U.S. preschools with less knowledge of the English language face in narrowing the achievement gap. However, all preschoolers' English vocabulary knowledge showed positive growth over the course of the year. Convergent with previous research (Mancilla-Martinez & Lesaux, 2011b), students designated as having less English language proficiency generally evidenced smaller English vocabularies at preschool entry. Despite this trend, the vocabulary differences were only large enough to distinguish the Non-English Proficient group from the English-Only group. Considering the demographic characteristics of the children in the sample, this finding is not entirely unexpected. All children attended preschool by meeting at-risk criteria, with nearly all residing in lowincome homes. The relationship between poverty and vocabulary attainment, regardless of children's language background, may explain the similarities between the children's vocabulary knowledge. Seminal work by Hart and Risley (1995) shows that children from lower-income homes tend to receive less language input that helps stimulate vocabulary learning, raising the possibility that the English language environment in terms of quality/quantity of use among the English Proficient group (which included the English-Only, Fluent English Proficient, and Limited English Proficient children) was more similar than different. Anecdotally, preschool administrators reported that even though children from English-Only homes are not administered the English language proficiency test, this has occurred and they often receive scores placing them at the Limited English Proficient level, raising the possibility that children attending state-funded preschool often enter with limitations at the language level. Our second finding revealed that all students evidenced English vocabulary growth over the course of the year, suggesting that the preschool experience might have served as a buffer. Schooling has been demonstrated to buffer the differences between children from different income backgrounds (Alexander, Entwisle, & Olson, 2007). But somewhat surprisingly, rates of vocabulary growth did not differ. This finding is especially concerning for the Non-English Proficient group given their depressed English vocabulary levels at preschool entry. By attending an English preschool program, one might expect that this group would evidence faster gains on English vocabulary (see MancillaMartinez & Lesaux, 2011b). In practical terms, the English Proficient group started out more than 6 months behind national monolingual English norms, but fell at the high-average range at the end of the year. Highlighting the larger vocabulary gap, the Non-English Proficient group started 18 months behind national norms, but nearly approached norms at the end of the year. Our third and final interrelated finding adds greater nuance and speaks to the value of moving beyond English language proficiency designation to understanding ELLs' developing vocabularies in a broader context. When cumulative risk was considered, differences in vocabulary were exacerbated for the Non-English Proficient group compared to the English Proficient group; despite having the same number of risk factors applicable to the family, children designated as less English language proficient were more adversely impacted in relation to their vocabulary attainment compared to their more English-proficient peers. This is concerning given that children in immigrant families, like many of the children in our study, are more likely to experience multiple risk factors critical to development compared to their non-immigrant peers (Hernandez, 2004). Individually, having single and limited English-speaking parents negatively related to preschoolers' English vocabulary levels. Past research shows that children growing up in single-parent homes (Cherlin, 1999; Fry & Scher, 1984) and in homes in which the majority language is not spoken proficiently (Hernandez, 2004; Suarez-Orozco & Suarez-Orozco, 2013) tend to evidence lower achievement outcomes.

85

Because children exposed more consistently to a language show advantages in their command of that language (Gathercole, 2002; Pearson, Fernández, Lewedeg, & Oller, 1997), it was unsurprising that the effect of having a limited English-speaking parent was magnified for the Non-English Proficient group compared to the English Proficient group. However, it was unexpected that having multiple families living in the home served as a protective factor. Past research, largely conducted with monolingual English-speakers, finds that multiple families in the home have negative effects on a wide range of outcomes, including academic development (e.g., Evans, Saegert, & Harris, 2001). Because having multiple families live in the home is a way for families to economize (Hernandez, 2004), it is not surprising that nearly half of all families in the sample reported doing so. Further research is needed to understand the mechanisms underlying the influence of multiple families in the home on student academic outcomes as we can only speculate that the additional household members may have served as language models for the children. 4.1. Implications In interpreting the results of this study, it is important to underscore that this sample of children had an average of 8 risk factors, which is not insignificant, especially for ELLs. Previous research shows that the more risk factors present in a child's life, the larger the disparities, with lowincome status emerging as the most common risk factor (Halle et al., 2009). In this study, nearly all families were low-income, and the presence of additional risk factors exacerbated the already salient impact of low-income status. At the same time, attention to the conceptualization of risk for this population of learners is warranted as one-third of the risk factors applied to 8% or less of the total sample. Of the remaining 20 at-risk factors, 9 applied to less than 20% of the full sample, effectively leaving 11 of the 29 atrisk factors with wide variability to consider. This is an important practical and policy point given that efficiency in identification of children as ‘at risk’ is a top priority for preschool programs. Our results suggest that the number of risk factors could be considerably focused for preschool screening purposes into categories of likelihood and consequent impact. Results of this study also call attention to the ongoing debates regarding the impact of various instructional models. Our results show that all English-instructed groups of children, regardless of English language proficiency designation, made gains in English vocabulary. However, concerns arise when these results are viewed as confirmation that native language instruction is unnecessary. This study was not designed to compare instructional models and thus cannot speak to the issue. However, quality of instruction is an active ingredient for language and literacy development (Barnett, Yarosz, Thomas, Jung, & Blanco, 2007; Burchinal, Field, López, Howes, & Pianta, 2012; Durán, Roseth, & Hoffman, 2010), necessitating that researchers, practitioners, and policymakers focus on better understanding the instructional context that can best support students' language learning. This is especially pressing for ELLs. As this study shows, children who enter preschool with lower English language proficiency levels not only start out lower on vocabulary knowledge, but also end the year at lower levels. 4.2. Future research As with any study focused on understanding a complex developmental process, the present study raises important questions to be addressed in future research. The dichotomously scored at-risk factors were taken from the preschool screener rather than having been selected a priori; a continuous classification could provide greater insight into how these factors matter, in what amount, and for whom. In a related vein, given the study's focus on potential differences in vocabulary outcomes by language proficiency, the Pre-IPT language proficiency designations were utilized. While there is no singular “best” approach to modeling language proficiency (Gee, Walsemann, & Takeuchi, 2010),

86

J. Mancilla-Martinez et al. / Learning and Individual Differences 35 (2014) 79–86

exploration of language proficiency as a continuous variable has the potential to provide additional insight. Additionally, we considered cumulative risk. While assuming that all of the risk factors hold equal weight, research with monolingual English speakers confirms the utility of this approach (e.g., Burchinal et al., 2000; Cadima et al., 2010) and we also explored the influence of individual risk factors. Furthermore, longitudinal work from preschool throughout the academic trajectory can illuminate whether risk factors vary in importance or predictive influence differentially across development. Given the impact of the risk factors addressed in this study on vocabulary development, the influences can be expected to continue to be relevant for other domains (e.g., reading). Similarly, longitudinal work can be informative for understanding factors that may predict learning outcomes (e.g., learning disabilities), helping with preventive intervention. We were also limited to administering the vocabulary measure in English given the district's practical and legitimate concern with additional testing time from preschoolers' instructional day. Yet, knowledge of students' Spanish vocabulary skills would have provided a more comprehensive understanding of their vocabulary knowledge. Likewise, given the study's focus, multilevel influences (e.g., classroom- and school-level) on vocabulary outcomes were not investigated, but their inclusion could provide added insight. Finally, while this study framed the influence of risk factors on vocabulary development, the findings can also be understood via a protective lens. Considering the nature of facilitating characteristics through a culturally sensitive lens, as highlighted in this study with the protective factor of having multiple families in the home, is necessary. The study findings make clear that English language proficiency designation and contextual factors play a mutually exacerbating role on vocabulary as early as preschool. Considering the significantly greater cost of remediation and intervention for struggling students, and the burden of secondary consequences (e.g., engaging in high-risk behaviors), this work lends additional support to the argument for early intervention for this growing population that addresses not only narrow dimensions of performance but considers a child in his or her context as a complex and dynamic system. References Alexander, K. L., Entwisle, D. R., & Olson, L. S. (2007). Lasting consequences of the summer learning gap. American Sociological Review, 72, 167–180. Anderson, R. C., & Freebody, P. (1983). Reading comprehension and the assessment and acquisition of word knowledge. Advances in reading/language research, 2, 231–256. August, D., & Shanahan, T. (Eds.). (2006). Developing literacy in second-language learners: Report of the National Literacy Panel-Minority Children and Youth. Mahwah, NJ: Erlbaum. Ballard, W. S., Tighe, P. L., & Dalton, E. F. (1979, 1989, 1991). Pre-IDEA oral language proficiency test, Oral English, ages 3–5. Brea, CA: Ballard & Tighe. Barnett, S. W., Yarosz, D. J., Thomas, J., Jung, K., & Blanco, D. (2007). Two-way and monolingual English immersion in preschool: An experimental comparison. Early Childhood Research Quarterly, 22, 277–293. Brooks-Gunn, J., & Duncan, G. (1997). The effects of poverty on children and youth. The Future of Children, 7, 55–71. Burchinal, M., Field, S., López, M. L., Howes, C., & Pianta, R. (2012). Instruction in Spanish in pre-kindergarten classrooms and child outcomes for English language learners. Early Childhood Research Quarterly, 27, 188–197. Burchinal, M. R., Roberts, J. E., Hooper, S., & Zeisel, S. A. (2000). Cumulative risk and early cognitive development: A comparison of statistical risk models. Developmental Psychology, 36(6), 793–807, http://dx.doi.org/10.1037//0012-1649.36.6.793. Burchinal, M., Vernon-Feagans, L., & Cox, M. (2008). Cumulative social risk, parenting, and infant development in rural low-income communities. Parenting: Science and Practice, 8, 41–69, http://dx.doi.org/10.1080/15295190701830672. Cadima, J., McWilliam, R. A., & Leal, R. (2010). Environmental risk factors and children's literacy skills during the transition to elementary school. International Journal of Behavioral Development, 34(1), 24–33, http://dx.doi.org/10.1177/0165025409345045. Cherlin, A. J. (1999). Going to extremes: Family structure, children's well-being, and social sciences. Demography, 36, 421–428. Cohen, J. (1992). A power primer. Psychological Bulletin, 112, 155–159. Conger, R. D., & Donnellan, M. B. (2007). An interactionalist perspective on the socioeconomic context of human development. Annual Review of Psychology, 58, 175–199. De Houwer, A. (1999). Two or more languages in early childhood: Some general points and practical recommendations. Washington, DC: Center for Applied Linguistics (Retrieved from www.cal.org/resources/digest/earlychild.html). Duncan, G. J., Ziol-Guest, K. M., & Kalil, A. (2010). Early childhood poverty and adult attainment, behavior, and health. Child Development, 81, 306–325.

Dunn, L. M., & Dunn, D.M. (2007). Peabody picture vocabulary test – 4. Bloomington, MN: NCS Pearson, Inc. Durán, L. K., Roseth, C. J., & Hoffman, P. (2010). An experimental study comparing English-only and transitional bilingual education on Spanish-speaking preschoolers' early literacy development. Early Childhood Research Quarterly, 25, 207–217. Evans, G. W., & Kim, P. (2010). Multiple risk exposure as a potential explanatory mechanism for the socioeconomic status-health gradient. Annals of the New York Academy of Sciences, 1186, 174–189, http://dx.doi.org/10.1111/j.1749-6632.2009.05336.x. Evans, G. W., Saegert, S., & Harris, R. (2001). Residential density and psychological health among children in low-income families. Environment and Behavior, 33, 165–180. Farah, M. J., Shera, D.M., Savage, J. H., Betancourt, L., Giannetta, J. M., Brodsky, N. L., & Hurt, H. (2006). Childhood poverty: Specific associations with neurocognitive development. Brain Research, 1110(1), 166–174. Felitti, V. J., Anda, R. F., Nordenberg, D., Williamson, D. F., Spitz, A.M., Edwards, V., & Marks, J. S. (1998). Relationship of childhood abuse and household dysfunction to many of the leading causes of death in adults: The Adverse Childhood Experiences (ACE) study. American Journal of Preventative Medicine, 14(4), 245–258. Fry, P.S., & Scher, A. (1984). The effects of father absence on children's achievement motivation, ego-strength, and locus-of-control orientation: A five-year longitudinal assessment. British Journal of Developmental Psychology, 2, 167–178. Fuller, B., Bridges, Bein, E., Jang, H., Jung, S., Rabe-Hesketh, & Kuo, A. (2009). The health and cognitive growth of Latino toddlers — At-risk or immigrant paradox? Maternal and Child Health Journal, 13, 755–768. García, E. E., & Frede, E. C. (Eds.). (2010). Young English language learners: Current research and emerging directions for practice and policy. New York, NY: Teachers College Press. García, E., & Jensen, B. (2009). Early educational opportunities for children of Hispanic origins. Social Policy Report, 23(2), 1–20. Garmezy, N., Masten, A. S., & Tellegen, A. (1984). The study of stress and competence in children: A building block for developmental psychopathology. Child Development, 55(1), 97–111. Gathercole, V. C. (2002). Monolingual and bilingual acquisition: Learning different treatments of that-trace phenomena in English and Spanish. In D. K. Oller, & R. E. Eilers (Eds.), Language and literacy in bilingual children. Multilingual Matters. (pp. 220–254). Gee, G. C., Walsemann, K. M., & Takeuchi, D. T. (2010). English proficiency and language preference: Testing the equivalence of two measures. American Journal of Public Health, 100, 563–569. Halle, T., Forry, N., Hair, E., Perper, K., Wandner, L., Wessel, J., & Vick, J. (2009). Disparities in early learning and development: Lessons from the Early Childhood Longitudinal Study – Birth Cohort (ECLS–B). Washington, DC: Child Trends. Hart, B., & Risley, T. R. (1995). Meaningful differences in the everyday experience of young American children. Baltimore: Brooks. Hernandez, D. J. (2004). Demographic change and the life circumstances of immigrant families. The Future of Children, 14, 16–47. López, M. H., & Velasco, G. (2011). The toll of the great recession: Childhood poverty among Hispanics sets record, leads nation. Retrieved from. http://www.pewhispanic.org/ 2011/09/28/childhood-poverty-among-hispanics-sets-record-leads-nation Mancilla-Martinez, J., & Lesaux, N. K. (2011a). Early home language use and later vocabulary development. Journal of Educational Psychology, 103, 535–546, http:// dx.doi.org/10.1037/a0023655. Mancilla-Martinez, J., & Lesaux, N. K. (2011b). The gap between Spanish speakers' word reading and word knowledge: A longitudinal study. Child Development, 82, 1544–1560, http://dx.doi.org/10.1111/j.1467-8624.2011.01633.x. Mancilla-Martinez, J., & Vagh, S. B. (2013). Growth in toddlers' Spanish, English, and conceptual vocabulary knowledge. Early Childhood Research Quarterly, 28, 555–567. National Task Force on Early Childhood Education for Hispanics (2007). Para nuestros niños: Expanding and improving early education for Hispanics—Main report. Tempe, AZ: National Task Force on Early Childhood Education for Hispanics (Retrieved from http://www.ecehispanic.org/work/expand_MainReport.pdf). NICHD Early Child Care Research Network (2004). Are child developmental outcomes related to before-and after-school care arrangements? Results from the NICHD study of early child care. Child Development, 75(1), 280–295. Pan, B.A., Rowe, M. L., Singer, J.D., & Snow, C. E. (2005). Maternal correlates of growth in toddler vocabulary production in low-income families. Child Development, 76(4), 763–782. Passel, J., Cohn, D., & Lopez, M. H. (2011). Hispanics account for more than half of nation's growth in past decade. Washington, DC: Pew Hispanic Center (http://pewhispanic. org/reports/report.php?ReportID=140). Pearson, B. Z., Fernández, S.C., Lewedeg, V., & Oller, D. K. (1997). The relation of input factors to lexical learning by bilingual infants. Applied Psycholinguistics, 18, 41–58. Rutter, M. (1979). Protective factors in children's responses to stress and disadvantage. In M. W. Kent, & J. E. Rolf (Eds.), Primary prevention of psychopathology. Social competence in children, vol. 3. (pp. 49–74). Hanover, NH: University of New England Press. Singer, J.D., & Willett, J. B. (2003). Applied longitudinal data analysis: Modeling change and event occurrence. New York, NY: Oxford University Press. Snow, C. (1992). Perspectives on second-language development: Implications for bilingual education. Educational Researcher, 21(2), 16–19. Snow, C. E., & Kim, Y. (2007). Large problem spaces: The challenge of vocabulary for English language learners. In R. K. Wagner, A. E. Muse, & K. R. Tannenbaum (Eds.), Vocabulary acquisition: Implications for reading comprehension (pp. 123–139). New York, NY: Guilford Press. Stanton-Chapman, T. L., Chapman, D. A., Kaiser, A. P., & Hancock, T. B. (2004). Cumulative risk and low-income children's language development. Topics in Early Childhood Special Education, 24(4), 227–237. Suarez-Orozco, C., & Suarez-Orozco, M. (2013). Conferring disadvantage: Immigration, schools, and the family. In D. Allen, & R. Reich (Eds.), Education, justice, & democracy (pp. 133–154). Chicago: The University of Chicago Press.