Oral language and reading abilities of f

Bryant, and Martie Skinner The University of North Carolina at Chapel Hill, USA T his study investigated the relationship between selected child and family demographic characteristics ch

http://www.tandf.co.uk/journals/pp/01650254.html DOI: 10.1080/01650250143000229

Oral language and reading abilities of �rst-grade Peruvian children: Associations with child and

family factors

Dina C Castro, Bobbie B Lubker, Donna M Bryant, and Martie Skinner

The University of North Carolina at Chapel Hill, USA

T his study investigated the relationship between selected child and family demographic characteristics (child age, child sex, child birth order, maternal education, and parent language status), family processes (parent-to-child reading at home, and parent expectations about child’s educational attainment), and preschool experience with poor Peruvian �rst-grade children’s oral language and reading abilities, and examined whether those factors help to explain differences among

children living in poverty First-grade students (N ˆ 137) of �ve schools in a poor neighbourhood of

Lima, Peru participated in the study Children were given picture vocabulary, verbal analogies, letter-word identi�cation, and reading comprehension tests Information about the children and their families was gathered through parent interviews Children whose parents had higher expectations obtained higher scores on picture vocabulary, verbal analogies, letter-word identi�cation and reading comprehension Children who attended private and public preschools obtained higher scores in letter-word identi�cation than those who did not attend preschool These �ndings support previous research on the relevance of family beliefs, above and beyond sociodemographic variables, as contributors to children’s oral language and reading, and provide some evidence of the bene�ts of preschool among children living in poverty Future research is recommended to identify the speci�c strategies used by low-income Peruvian parents with high expectations to support their children’s language and reading; and to determine the relationships between type and quality in Peruvian preschool programmes, and programme practices that may differentially affect children’s language and reading skills.

Introduction Extensive poverty is a major barrier for child development and

education in Peru Sixty per cent of Peruvian children, from

birth to 17 years of age, live in poverty (Webb & Ferna´ndez

Baca, 1999) In 1993, only 46% of children 12 years of age and

younger lived in homes that had basic living amenities (potable

water, sewer, and electricity) (Webb & Ferna´ndez Baca, 1995)

Economic crises and political violence accelerated

rural-to-urban migration during recent decades Currently, 72% of

Peruvians live in urban areas Typically, migrants start new

communities around the cities called ‘‘asentamientos humanos’’,

where they live in extreme poverty

Peruvian school systems are challenged to serve this

population The percentage of school enrolment for Peruvian

children 6–11 years of age has been increasing; in 1998 it was

93% However, the percentage of grade retention in �rst grade

has been around 30% for many years Failure in learning to

read and write was the main reason to retain children in �rst

grade In 1995, as part of the implementation of the

Articulation of Preschool and Primary Education Programme,

Peruvian governmental of�cials decided that all �rst-grade students in public education would be promoted to second grade That gave them another year to complete their reading and writing learning processes This decision is re�ected in the country’s educational statistics as a notable decrease in grade retention In 1998, grade retention in second grade was 17% (Webb & Ferna´ndez Baca, 1999) Considering that these are national data, it is possible that these percentages may be higher in poor urban and rural areas Poor urban children from

the asentamientos humanos in Peru are at increased risk for

school failure, which is manifest in the early grades by dif�culties in developing oral language, reading, and writing abilities

Research in developing countries faces methodological challenges mainly related to the measurement instruments used The majority of psychoeducational instruments available

in Peru are translations from instruments developed in the USA and other industrialised countries The present study is not exempted from these challenges, and is an effort to generate sound research data that truly re�ect the character-istics of Peruvian children and families

Correspondence should be addressed to Dina C Castro, Frank Porter

Graham Child Development Center, The University of North

Carolina at Chapel Hill, 105 Smith Level Rd., CB# 8180, Chapel

Hill, NC 27599-81 80, USA (E-mail: Dina_Castro@unc.edu).

This study is based on Dina C Castro’s doctoral dissertation The

authors would like to thank Drs Margaret Burchinal, and Kinnard

White for their advice on an earlier version of this manuscript, Jose´ Miguel Sandoval for his assistance in data management and analysis, Edith Noriega for her assistance in collecting data, and especially the children, parents, school administrators and teachers in Lima, Peru for their kind participation in this study.

Language development of children reared in poverty

Several investigators have shown that the compounding of risk

conditions in poverty environments has negative in�uences on

children’s physical, mental, and social development (e.g.,

Margolis, Greenberg, & Keyes, 1992; Ricciuti & Scarr, 1990;

Sameroff, Seifer, Barocas, Zax, & Greenspan, 1987; Wasik,

Ramey, Bryant, & Sparling, 1990) ‘‘The chronic stress and

diminished material and psychological resources that often

characterize poverty environments combine in synergistic

fashion to the detriment of young children’’ (Bradley et al.,

1994, p 347) Speci�cally, the language developmen t of

children living in poverty is below what is expected for their

chronological age (Montenegro, 1992) Social class differences

have been found in children’s oral language skills (Dickinson &

Snow, 1987) and in children’s concepts of printed language

(Ferreiro & Teberosky, 1982)

Evidence from research indicates that language

develop-ment is associated with school achievedevelop-ment, especially with

literacy-related skills Low level of skills in oral language during

preschool years are often associated with later reading

dif�culties, especially for children from low SES families

(Raz & Bryant, 1990; Scarborough, 1990) Walker,

Green-wood, Hart, and Carta (1994) reported the results of a 10-year

longitudinal study, examining associations between early

language development and family SES Children’s assessments

before 3 years of age were compared with later measures of

language and academic achievement in kindergarten through

third grade The results showed that differences in child

language before schooling, related to family SES, were

predictive of children’s scores in verbal ability, receptive and

spoken language, and academic achievement (reading,

lan-guage expression, and math) tests from kindergarten through

third grade

Early childhood education as a protective factor

Internationally, a large body of research has demonstrated that

early childhood education can prevent early academic failure

and that it has positive long-lasting effects throughout the lives

of children from poor families (e.g Boocock, 1995; Burchinal,

Campbell, Bryant, Wasik, & Ramey, 1997; Myers, 1992)

High-quality early educational intervention (EEI) may

ameli-orate and prevent further deterioration and delays in children’s

development, particularly for children who live in poverty

(Consortium of Longitudinal Studies, 1983) During recent

decades, the importance of early childhood education

pro-grammes has been increasingly recognised at academic and

political levels in countries around the world (Woodill,

Bernhard, & Prochner, 1992) The care and education of

young children are signi�cant issues because of the increased

proportions of children living in poverty and their

over-representation in special education (Hauser-Cram, Pierson,

Walker, & Tivnan, 1991)

It is necessary to emphasise the preventive perspective

underlying EEI as an effort to counteract the effects of poverty

for children (Martin, Ramey, & Ramey, 1990) As mentioned

earlier, the conditions of poverty constitute risk factors for

children’s early cognitive and social development that are

frequently associated with later school failure (Wasik et al.,

1990) High-quality, family-oriented, comprehensive EEI may

serve as a primary prevention strategy, since it may ensure that

these children enter school as healthy and competent learners,

which will in�uence their subsequent success in school, thus improving their prospects for a more productive and personally satisfying life (Bryant & Maxwell, 1997; Schweinhart, Barnes,

& Weikart, with Barnett & Epstein, 1993)

Studies on quality in early childhood care and education have demonstrated the relationships among quality indicators such as class size, adult:child ratios, and teacher education and children’s cognitive, language, and social developmen t (Burch-inal, Roberts, Nabors, & Bryant, 1996; Burchinal et al., 2000; Howes, Phillips & Whitebook, 1992) Teachers who have smaller class sizes and children to adult ratios are more likely to interact positively and sensitively to the children in their classrooms (Phillipsen, Burchinal, Howes, & Cryer, 1997) Similarly, more educated teachers are more likely to have classes rated to be of higher quality (Whitebook, Howes, & Phillips, 1989)

In Peru, 52% of children 3–5 years of age receive preschool education (Webb & Ferna´ndez Baca, 1999), which includes private and public preschools and nonformal preschools Private and public preschools are generally conducted by certi�ed preschool teachers who have nonprofessional teacher assistants These preschools usually have adequate buildings and appropriate equipment and teaching materials

The largest proportion of preschools available in poor urban and rural areas are the ‘‘PRONOEI’’, a nonformal modality of preschool developed to expand preschool coverage for children

in poverty, at a lower cost to the government Nonformal preschools are conducted by paraprofessionals, who usually have no more than high school education and who are paid less than the minimum wage These persons participate in a short initial training programme and periodic refresher training Preschool teachers assigned by the school districts provide supervision Many of these preschools operate in inappropri-ate, sometimes improvised facilities, and lack equipment and materials

Studies of preschool education in Peru and its impact on poor children’s oral language and reading have not been found Based on international literature indicating the bene�ts of preschool attendance and its differential effects related to type and quality of the preschool programme, in this study we hypothesised that on measures of language and reading (a) �rst graders who attended preschool will perform better than those who do not, and (b) �rst graders who attended private and public preschools will perform better than those who attended nonformal preschools

Family factors and child language and literacy development

There are important differences among families living in poverty, some of them critical for children’s language and literacy developmen t and school success Differences among low income families may not only be related to demographic family characteristics, such as parental education or family composition, but also to family process factors such as parental beliefs and behaviours In research on child development and education, considerably more attention has been given to family demographic characteristics than to family processes when studying differences in child outcomes among children from low-income families The literature on developmen t of children in poverty and resilience shows an increase in research focusing on those family processes, such as parental educa-tional expectations and reading practices at home, that may

help some disadvantaged children develop into

well-function-ing individuals and to succeed in school, in spite of the poverty

situation in their homes (e.g., Bradley et al., 1994; Halle,

Kurtz-Costes, & Mahoney, 1997; Zimmerman, & Arunkumar,

1994)

Research �ndings indicate that family factors related to

language and literacy developmen t differ in home

environ-ments of poor versus nonpoor young children (Bradley et al.,

1994) Fewer studies have investigated differences among

families within the low SES group and their relation with

children’s language and reading abilities Among those studies

some have focused on family process variables such as parental

expectations and parent behaviours Parent expectations about

their children’s educational attainment have been shown to be

signi�cantly associated with children’s school achievement

(i.e., reading and math scores) in the USA (Halle et al., 1997;

Mau, 1997) and in other countries (Morgan, Shiel, Hickey, &

Forde, 1995; Phillips, 1992) For the Peruvian population,

speci�cally for poor children and families, studies on the

relationships between family factors and child language and

literacy developmen t are scarce One study was found that

investigated family factors related to child reading and

mathematics achievement in �rst-grade Peruvian children

from indigenous Quechua-speakin g families, living in three

different geographic locations Parental expectations about the

child’s future work and the parental behaviours of helping with

homework and teaching style were found to be positively

related to �rst-graders’ scores on reading and math tests

(Barber, 1988)

Regarding parent behaviours, research �ndings are not

consistent For example, some studies have found substantial

differences in literacy-related behaviours among low-income

families and those differences were strongly associated with

children’s oral language skills (Payne, Whitehurst, & Angell,

1994; Ricciuti, White, & Fraser, 1993) Contrarily, another

study found no signi�cant relationship between the amount of

exposure to reading in the home and the language skills of

preschool children (DeBaryshe, Rodarmel, Daly, & Huntley,

1992) For this study, we hypothesised that children whose

parents have higher expectations about their educational

attainment and who read to them more often and for longer

periods of time will obtain higher scores on measures of oral

language and reading abilities

The purposes of the study presented here were to investigate

the relationship between selected child and family

demo-graphic characteristics, family processes, and preschool

experi-ence with poor Peruvian �rst-grade children’s oral language

and reading abilities and to determine whether those variables

help to explain differences among children living in poverty

Methods

Participants and procedures

The study was conducted with 137 �rst-grade Peruvian

Spanish-speaking children and their parents Children were

randomly selected among �rst-grade students at �ve public

schools located in a poor urban neighbourhood in Lima, Peru

All �rst-grade classes at each of the 5 schools (17 in total) were

included in the study, except for 2 classes of repeaters in one

school The number of students per class ranged from 33 to

44 First-grade children who had not repeated �rst grade and

who had no cognitive, physical, and/or psychological disabil-ities, as reported by teachers, were included Siblings were not included After exclusions, a 25% random sample of children per classroom yielded an overall sample of 150 children From the 150 initial sample, 142 parents were interviewed By the time of the assessment 5 children of the interviewed parents were no longer attending school Thus, from the 142 parents interviewed, 137 children were assessed Data from parent interviews without the corresponding child assessment data were not included Parents were interviewed at their homes or

at the school before starting children’s assessments Children were assessed 1 month before the school year ended The �rst author and two research assistants, who were Peruvian licensed psychologists, conducted data collection

Child and family predictor variables

For this study child and family predictor variables were obtained from the parent interviews Child and family demographic variables were: child age (in months), child sex, child birth order (in two strata: �rstborn, and birth order 2–9), maternal education (mother’s years of schooling), and parent language status (score on an 8-item scale) Family process variables were: parent-to-child reading (score on a 4-item scale) and parent expectations about child’s educational attainment (high school or less, technical, and university) These variables represent the response of the parent who was interviewed Parent respondents were 118 mothers and 19 fathers Type of preschool the child attended (no preschool, nonformal preschool, and public/private preschool) was also a predictor variable in this study The parent language status scale contains questions about both parents’ Spanish language status, as native or second language, and the age when and the way in which they learned Spanish Values ranged from 5

to 28, with a higher score indicating that a parent learned Spanish in childhood, and at school The parent-to-child reading scale contains questions that included frequency, reading time per session, and type of reading Values ranged from 0 to 14, with a higher score indicating more frequent and complex parent-to-child readings Table 1 presents data

on sample characteristics

Child outcome variables

The outcome variables were children’s scores on two measures

of oral language, Picture Vocabulary and Verbal Analogies, and two measures of reading abilities, Letter-Word Identi�ca-tion (LWI) and Reading Comprehension Children were individually given three tests of the Spanish version of the Woodcock-Mun˜oz Language Survey (WMLS-S; Woodcock & Mun˜oz-Sandoval, 1993): Picture Vocabulary, Verbal Analo-gies, and Letter-Word Identi�cation The measure of reading

comprehension was the Prueba de ComprensioÂn Lectora de

Complejidad Lingu¨õÂstica Progresiva (CLP; Alliende,

Conde-mar´n, & Milicic, 1993), given to groups of 10–12 children.õ

Table 1 shows descriptive data on these measures

Other descriptive demographic information

Contextual information about the children and families who participated in the study was also gathered through the parent interviews Among fathers, 52% had studied at high school (complete or incomplete) All fathers in the sample were

employed; however, the majority of them were employed in

menial jobs and jobs with unstable low incomes (e.g., street

vendor, construction worker, driver) This was also the case for

most of the 44% of mothers who were employed Most of the

parents had migrated from rural Andean communities, and

around 60% of them spoke Spanish as a second language

Another characteristic of families in this study is that both

parents lived in the homes of almost 80% of the families

Measures

Woodcock-Mun˜oz Language Survey (WMLS-S), Spanish

ver-sion. The WMLS-S is strati�ed from preschool through

adulthood and by grade levels It contains four tests: Picture

Vocabulary, Verbal Analogies, Letter-Word Identi�cation, and

Dictation Dictation was not included in this study Picture

Vocabulary measures the capability to name pictured objects

It is an expressive semantic task at the single-word level, which

gradually presents familiar, then less familiar vocabulary

Verbal Analogies measures the ability to understand and

complete a logical word relationship in oral language The

vocabulary remains relatively simple, but relationships among

words become increasingly complex This is a task of reasoning

with the language In Letter-Word Identi�cation, the sample

and �rst four items measure the capability to match a

pictographic representation of a word with a picture of an

object The child is asked to put his/her �nger on the drawing

that represents the big picture The remaining items measure the child’s ability to recognise individual letters and words It is not assumed that the child knows the meaning of any word The items become more dif�cult as they are comprised of words that appear less and less often in written language (Woodcock & Mun˜oz-Sandoval, 1993)

The test authors provide norms for the WMLS-S by equating levels of Spanish language performance to similar levels of language performance in English The equating normative data were from more than 2000 Spanish-speaking persons from Argentina, Costa Rica, Mexico, Peru, Puerto Rico, Spain, and the United States Internal consistency reliability coef�cients are reported by the authors only for the English version of the test Those test reliabilities were calculated by the split-half procedure by age groups Reported reliability coef�cients for 6-year-olds were: Picture Vocabu-lary 77, Verbal Analogies 81, and Letter-Word Identi�cation 96

Prueba de ComprensioÂn Lectora de Complejidad Lingu¨õÂstica Progresiva (CLP). This standardised test was developed in Chile to assess reading comprehension of Spanish-speaking children The CLP includes eight reading levels, becoming increasingly dif�cult from linguistic and comprehension perspectives Each level has a parallel form The areas of reading assessed in �rst graders are word, phrase, and sentence comprehension with tasks requiring the children to draw a line

Table 1

Sample description: Children and family characteristics ( N ˆ 137)

Children

Birth order

Type of preschool attended

Oral language scores:

Reading scores a

Family

Parent expectations about child’s educational attainment

a For descriptive reasons the reading scores are presented here in their raw score form In the statistical analyses these scores were dichotomised.

to match words, phrases, or sentences with corresponding

pictures Six subtests of the �rst level of reading

comprehen-sion (three from Form A and three from Form B) were

administered A total reading comprehension score was

obtained by adding the correct responses from the six subtests

Normative data for this test were from a sample of 641 Chilean

subjects strati�ed by sex, grade level, and socioeconomic

status For the �rst �ve levels of reading comprehension,

test-retest reliability estimates over 3 weeks were obtained from a

subsample of 44 children Those reliability coef�cients were

.97 for Form A and 90 for Form B

Parent interview. The parent interview protocol included

demographic items about the child and family and questions

about the child’s preschool experience, parent expectations

about her/his child’s education, and home literacy

environ-ment

Pilot phase of the study

A pilot study was conducted with 15 Peruvian children in �rst

grade who were not included in the study sample This pilot

testing of the complete battery revealed that: (1) Children in

this sample were not familiar with some of the objects

represented in the picture vocabulary test Most or all children

failed three items of the picture vocabulary test of the

WMLS-S (items were a cheque, a tennis racquet, and a postage stamp)

(2) Children gave wrong answers or did not respond to most

items in test 3 of Form B of the reading comprehension test

Pictures of activities common in rural areas were not familiar to

children living in a city, for example, a horse pulling a plough

Also, �ve of seven items in test 3 of Form B of the reading

comprehension test included words not commonly used in the

Spanish spoken in Lima or words with different meanings For

instance, in the sentence ‘‘A Moro lo amarraron a un poste’’

(Moro was tied to a stick), the word poste is used to refer to a

‘‘stick’’; however, in urban areas of Lima, the word poste is used

to name an electricity pole As a result of the pilot study, test 3

of Form B was not used The picture vocabulary test of the

WMLS-S was not changed

Reliability and validity

Reliability analyses were carried out on parent language status

and parent-to-child reading scores The Cronbach’s alpha

reliability coef�cient for the 8-item scale of parent language

status and for the 4-item scale of parent-to-child reading was

.88 Reliability analyses were also carried out on children’s

WMLS-S and CLP scores The reliability coef�cient obtained

for Picture Vocabulary was low (.41), probably because a large

number of items had no variance The score options in this test

were incorrect or correct Most children in the sample

answered many items correctly and many other items

incorrectly To obtain a higher reliability coef�cient, the �ve

most discriminating items in this test were selected for the

statistical analyses, i.e., items with around 50% of correct

responses The reliability coef�cient for these �ve items from

Picture Vocabulary was 71 Other reliability coef�cients were

.63 for Verbal Analogies, 97 for Letter-Word Identi�cation,

and 98 for Reading Comprehension The correlation between

the letter-word identi�cation test of the WMLS-S and the CLP

reading comprehension test (.88) was used as a measure of

predictive validity

Data analyses

The �rst step in the analyses was to examine descriptively the associations among predictor and outcome variables, for which correlations and ANOVAs were performed Next, the relations between child and family variables and children’s scores on Picture Vocabulary and Verbal Analogies were explored by performing a hierarchical multivariate multiple regression with subsequent univariate analyses In this analysis two groups of predictor variable values were entered as follows: �rst, the variables indicating child and family demographic character-istics (child sex, child age in months, child birth order, maternal education, and parent language status), and second, the family process variables (parent-to-child reading at home, and parent expectations about child’s educational attainment) and type of preschool the child attended

Two hierarchical logistic regressions were performed to explore the relations between child and family variables and children’s scores on Letter-Word Identi�cation (LWI) and Reading Comprehension, respectively Logistic regressions were necessary because preliminary descriptive data analyses showed non-normal distributions of the scores on the reading tests To meet the requirement of logistic regression, children’s scores in the two reading tests were dichotomised into two categories: lower and higher dif�culty Scores were dichot-omised using levels of test item dif�culty as criterion For LWI, scores in the lower dif�culty category correspond to children who could identify only letters and one-syllable words Scores

in the higher dif�culty category correspond to children who could also identify words of two and more syllables For reading comprehension, scores in the lower dif�culty category correspond to children’s reading comprehension ability at the one-word level Scores in the higher dif�culty category correspond to children’s reading comprehension ability at the sentence level For both analyses, variables were entered in two steps, in the same order as for the oral language analysis

Results

Correlations among predictor variables and children’s language and reading abilities

Descriptive analyses were conducted to examine the associa-tion among predictor variables, and children’s scores on the oral language and reading measures As can be seen in Table 2, these analyses show modest to moderate signi�cant positive correlations among maternal education, parents’ language, parent-to-child reading at home, parent expectations about child’s educational attainment, and type of preschool child attended One-way ANOVAs performed on the categorical predictor variables indicate that child sex was not signi�cantly related to any of the other (continuous/ordinal) predictor variables, and child birth order had a signi�cant positive relationship with maternal education (Eta2

ˆ 03; p <.05) and

parent-to-child reading at home (Eta2

ˆ 04; p < 05).

Regarding bivariate relationships between predictor and outcome variables, for the language measures, child sex and parent expectations were signi�cantly associated with Picture Vocabulary scores; and parent expectations was the only predictor variable associated with Verbal Analogies scores (Table 3) For the reading measures, child birth order, parent-to-child reading, parent expectations, and type of preschool were signi�cantly associated with Letter-Word Identi�cation

and Reading Comprehension scores In addition, child age was

associated with Reading Comprehension scores (Table 4)

Which child and family variables were signi�cantly

related to children’s oral language scores?

There was no statistically signi�cant overall relationship, at the

multivariate level, between the child and family demographic

variables as a group and the two measures of children’s oral

language However, as shown in Table 3, there was a

signi�cant overall relationship between child sex and Picture

Vocabulary and Verbal Analogies, multivariate F(2, 127) ˆ

3.37; p <.05 This result seems to be primarily attributable to

the relationship between child sex and children’s Picture

Vocabulary scores Child sex accounted for 5% of the variance

in performance on Picture Vocabulary, with boys (M ˆ 3.0)

scoring signi�cantly higher than girls (M ˆ 2.2).

There was no statistically signi�cant overall relationship

among the predictor variables as a group and children’s scores

in Picture Vocabulary and Verbal Analogies when the family

process and preschool variables were entered, after controlling

for child and family demographic characteristics However,

when this second group of variables was entered, there

continued to be a signi�cant overall relationship between child

sex and children’s oral language Also, there was an overall

signi�cant relationship between parent expectations about child’s educational attainment, and children’s oral language,

multivariate F(2, 127) ˆ 4.65; p <.05 These results appear to

be attributable to the relationship between parent expectations about child’s educational attainment and children’s scores on both Picture Vocabulary and Verbal Analogies The parent expectations variable accounted for 24% of the variance on Picture Vocabulary scores and 26% of the variance on Verbal Analogies Results of the multivariate multiple regression are presented in Table 3

A Bonferroni post hoc test with a signi�cance level set at 05 was performed to identify the signi�cant differences among the three subcategories of parent expectations about child’s educational attainment for each oral language measure Among the three levels of expectations, children whose parents

had the highest level of expectation (university, M ˆ 3.0)

obtained signi�cantly higher scores in picture vocabulary than children whose parents had lower expectations (high school or

less, M ˆ 1.9) Similarly, children whose parents had the highest level of expectation (university, M ˆ 4.4) obtained

signi�cantly higher scores in verbal analogies than children

whose parents had lower expectations (high school or less, M ˆ

2.8) There were no signi�cant differences in children’s oral language scores between the ‘‘technical’’ group and the

‘‘university’’ or ‘‘high school or less’’ groups

Table 2

Spearman’s correlations among continuous and ordinal predictor variables

Predictor variables

Maternal education

Parent language status

Parent-to-child reading

Parent expectations

Type of preschool attended

* p<.05; ** p<.01.

Table 3

Results of hierarchical multivariate multiple regression on two measures of oral language of first-grade Peruvian children ( N ˆ 137)

Univariate Picture Vocabulary Verbal Analogies Multivariate

Step 2: Family process and preschool variables

7 Parent expectations about child’s

educational attainment

* p<.05; ** p<.01.

Which child and family variables were signi�cantly

related to children’s letter-word identi�cation and

reading comprehension scores?

Results of the logistic regression indicate that the group of child

and family demographic variables did not have an overall

statistically signi�cant relationship with scores in Letter-Word

Identi�cation (LWI) However, child birth order was

signi�-cantly associated with children’s scores on LWI When the

family process and preschool variables were entered, the model

had an overall statistically signi�cant relationship with scores

on LWI, x2

ˆ 17.71; p < 001 Speci�cally, there were

signi�cant relationships between parent expectations and type

of preschool child attended with children’s scores on LWI The

birth order variable was no longer signi�cant (Table 4) Post

hoc tests of signi�cant differences in proportions set at a 05

level of signi�cance were performed to determine if differences

between groups in the parent expectations and type of

preschool variables were signi�cant Children whose parents

had the highest expectations about their children’s educational

attainment (university) were more likely to be in the higher

dif�culty category of LWI than were those whose parents had

lower expectations (technical and high school or less) The

difference between the ‘‘technical’’ and ‘‘high school or less’’

groups was not signi�cant Parent expectations accounted for

8% of the variance on LWI scores Similarly, children who

attended private or public preschools were more likely to be in

the higher dif�culty category of LWI than those who did not

attend preschool There was no signi�cant difference in

children’s LWI scores between children attending nonformal

preschools and either those who attended public/private

preschools or did not attend preschool Type of preschool

attended accounted for 7% of the variance on LWI

A signi�cant relationship was found between child and

family demographic variables and children’s Reading

Com-prehension scores, x2

ˆ 16.99; p < 01 As shown in Table 4,

speci�cally, child age and child birth order were signi�cantly

related to children’s performance on the reading

comprehen-sion test Older children were more likely to be in the higher

dif�culty category of Reading Comprehension than younger

children Child age accounted for 4% of the variance on

Reading Comprehension scores Children who were born �rst

in the family were more likely to be in the higher dif�culty

category of Reading Comprehension than those who had a higher birth order Child birth order accounted for 8% of the variance on Reading Comprehension scores

When the family process and preschool variables were entered, the model was signi�cantly related to children’s Reading Comprehension scores, x2

ˆ 22.40; p < 0001 Child

age and birth order continued to be signi�cant In addition, the parent expectations variable was signi�cantly associated with children’s Reading Comprehension scores (see Table 4) The post hoc test of signi�cant difference in proportions was performed to determine whether the differences between groups in the parent expectations variable were signi�cant Children whose parents had the highest expectation about their children’s educational attainment (university) were more likely to be in the higher dif�culty category of Reading Comprehension than were those whose parents had lower expectations (technical or high school or less) The difference between the ‘‘technical’’ and ‘‘high school or less’’ groups was not signi�cant Parent expectations accounted for 11% of the variance on Reading Comprehension scores The percentage of children in the high dif�culty category of LWI and Reading Comprehension and the results of the post hoc tests are presented in Table 5

Discussion The results of this study provide further evidence of the importance of family beliefs and behaviours and the bene�ts of preschool education on disadvantaged children’s development Parent expectations about their child’s educational attainment emerged as the strongest variable across measures of children’s oral language and reading, after child and family demographic variables were controlled One interpretation of the importance

of this variable is that in the context of this study population with relative homogeneity in socioeconomic status, the expectations of parents regarding their children’s educational attainment may be a proxy for several crucial aspects of parent behaviours for enhancing children’s performances in oral language and reading, including communicating their high expectations to their children Thus, another way in which the parent expectations variable may in�uence children’s

perfor-Table 4

Results of hierarchical logistic regression on two measures of reading abilities of first-grade Peruvian children ( N ˆ 137)

Letter-Word Identi�cation Reading Comprehension

Step 1: Child and family demographic variables

Step 2: Family process and preschool variables

7 Parent expectations about child’s

educational attainment

* p<.05; ** p<.01; *** p<.001.

mance is through children’s efforts to ful�l their parents’ high

expectations High parental expectations may be explored as

an important factor that promotes resiliency and increases

child motivation

The importance of parental expectation is emerging in the

international literature A positive relationship between parent

expectations and children’s reading and math scores was found

by Phillips (1992) in a study of the in�uence of parental

expectations and short-term goals on the school achievement

of children from second through sixth grade in Canadian

schools Similarly, parental expectation was found to be a

signi�cant predictor of reading achievement in a national study

conducted with a sample of over 3000 �fth-grade students in

the Republic of Ireland (Morgan et al., 1995) From a different

perspective, Australians Carpenter and Fleishman (1987)

found that as the level of parent expectations about their

children’s education increased, students’ achievement in

higher education also increased

In the USA parents’ educational expectations have been

found to be positively associated to children’s achievement in

different racial/ethnic groups For example, Halle et al (1997)

reported positive and signi�cant relationships between parent

expectations and children’s achievement in reading and math

in a group of African-American students in third and fourth

grades Another study with low-income 12-year-old

African-American children found that parent expectations of

educa-tional attainment consistently predicted those children’s

academic performance (Reynolds, 1998) In a study that

compared Asian immigrant, Asian American, and White

American high school students (Mau, 1997), it was found

that high parental expectations seem to be associated with

Asian students’ academic success regardless of their status as

new or established immigrants

Although this study did not measure children’s perceptions

of their parents’ expectations, other studies have, Goyette and

Xie (1999) found that high parental expectations explained a

large portion of children’s higher educational expectations

among Asian-American students participating in the study

compared with White students In a study with low-income African-American children in sixth grade, it was found that children’s perceptions of parent and teacher expectations were associated with reading and math achievement, above and beyond sociodemographic variables and measures of parent and teacher expectations (Gill, 1997)

From the evidence provided in the literature and our study

�ndings, questions about parental expectations that may be useful to explore in future research are: What strategies do Peruvian parents who have high educational expectations for their children use to support their children’s developmen t and learning? Is children’s better performance in school the result

of their efforts to ful�l their parents’ high expectations? As children perform better in school, do parental expectations increase? Does this in turn provide further motivation for children to improve their performance at school?

Our �ndings related to family processes associated with children’s language and reading skills indicate that future studies with the Peruvian population should explore the role of parental beliefs and behaviours in promoting child language and literacy development and school success among children in poverty The strong effect of the parent expectations variable found in this study may indicate that this variable contains information about a wide range of effective strategies that parents with high expectations use and that have not been measured in this study For example, parents with high expectations may provide their children with an affective environment at home, structure for learning, and use discipline and organisation more effectively than parents with lower expectations Future research should explore these and other strategies In this study, parent expectations were correlated with the type of preschool their children attended, thus, it seems appropriate to suggest that choosing a particular type of preschool for their children might be another example of a strategy or action that would support child developmen t and education

Regarding preschool experience, the type of preschool attended was found to be signi�cantly related to children’s

Table 5

Percentage of first-grade Peruvian children in high difficulty group of Letter-Word Identification and Reading

Comprehension by subcategories of significant predictor variables ( N ˆ 137)

Letter-Word Identi�cation Reading Comprehension

Predictor variables n

% children

in high groupa

Between groups difference

Extreme groups difference

% children

in high groupa

Between groups difference

Extreme groups difference

1 Child birth order

2 Parent expectations about

child’s educational attainment

3 Type of preschool attended

a Percentage of children in high dif�culty group.

* p<.05 n.s ˆ no signi�cant predictor variable.

scores on letter-word identi�cation In general, this �nding in

this population is con�rmatory of other research �ndings in

countries around the world The positive effects of preschool

education on poor children’s developmen t and school

achieve-ment are well-docuachieve-mented (e.g., Boocock, 1995; Campbell &

Ramey, 1994; Schweinhart et al., 1993)

The differences found in children’s scores on the letter and

word identi�cation measure according to the type of preschool

they attended suggest that we must formulate more speci�c,

detailed research questions about the relationship between type

and quality of preschool services in Peru Results of this study

indicate that children attending private or public preschools

obtained signi�cantly higher scores than children who did not

attend preschool One explanation of this �nding may be the

lack of access to written materials in the homes of low-income

families, thus, attendance at preschool may give children in

these families an opportunity to become familiar with written

materials and have experiences that support literacy

develop-ment

The letter and word identi�cation scores of children

attending nonformal preschools were lower than those of

children who attended private or public preschools, and higher

than those of children who did not attend preschool; however,

those differences were not signi�cant This could be

inter-preted as a difference in the amount, intensity, and/or quality

of the literacy experiences offered by nonformal preschool

versus those offered by a private or public preschool It may be

that what these programmes offer to the children is more than

what they receive at home, but not enough to make a

signi�cant impact in their reading skills Differential effects

on children’s development have been reported from other

studies conducted with children attending different types of

day care services and preschools and according to the quality of

services provided (e.g., Burchinal, Lee, & Ramey, 1989;

Burchinal et al., 1996; Peisner-Feinberg & Burchinal, 1995;

Schliecker, White, & Jacobs, 1991) This study does not

provide enough evidence of possible differential effects by type

of preschool service on Peruvian children’s reading abilities,

and speci�cally not about the impact of nonformal preschools

Considering that nonformal preschools are the most common

preschool service available for low-income populations in Peru,

further research needs to be conducted to determine the

characteristics and the extent to which this type of preschool

may affect Peruvian children’s developmen t and education In

general, it seems necessary to gain objective information about

the characteristics of the preschool services poor children are

receiving and to �nd ways to measure quality in this culture, if

we want to inform policy and practice on the factors in

preschools that increase Peruvian children’s probabilities of

school success

In contrast to �ndings in USA studies (Bryant,

Peisner-Feinberg, & Clifford, 1993; Peisner-Feinberg & Burchinal,

1995), in the present study no signi�cant relationship was

found between type of preschool and children’s oral language

scores This raises further questions Why was type of

preschool signi�cant for a reading measure but not for oral

language in Peruvian children? What characteristics in each

preschool type are associated with reading and oral language in

this population? In the statistical analysis, the relationship

between type of preschool and oral language measures

approached signi�cance; however, this may indicate that the

preschool programmes attended by children in the study—

even those presumed to provide better educational services—

were not good enough to signi�cantly affect children’s oral language scores Classroom experiences that support language development are characterised by the effective use of materials, activities, and teaching interactions to help children learn to communicate in words and to use learning concepts Factors such as a high adult:child ratio, lack of educational equipment and materials, and/or lack of preparation of teachers, may prevent preschool programmes from accomplishing their goals regarding children’s language development These �ndings suggest the need for further investigation of Peruvian pre-schools, including measures of a variety of dimensions of quality and their relationship with speci�c child developmental outcomes

The lack of signi�cance of the variables maternal education, parent-to-child reading, and parent language status are also interesting �ndings of this study As shown in Table 2, these three variables are correlated with the parent expectations variable and with each other As explained earlier, in this study parent expectations may represent an array of parent behav-iours, in addition to parent-to-child reading, which have not been measured The over-riding importance of parental expectations in this population seems to be in�uencing the contributions of other variables Finally, the nonsigni�cance of the parent language status variable indicates that it may be more important to examine the amount and quality of the linguistic interactions at home rather than parent language status This is related to our perspective about the relevance of studying the role of family processes instead of, or in addition

to, family demographics to identify differences among low-income families that may be associated with child develop-ment

One limitation should be considered when interpreting the results of this study regarding children’s Picture Vocabulary scores Dif�culties in obtaining acceptable reliability estimates for this subtest lead to our use of a small number of items in the analysis, thus, generalisability of these results should be viewed with caution

Issues on cross-cultural testing

The challenges experienced in this study with instruments for measuring oral language and reading abilities con�rm the need for developing culture-speci�c instruments for Spanish-speak-ing populations The extent to which a translated language test

is appropriate for use in a different cultural and linguistic setting seems questionable An example from this study is the

item ‘‘llama’’ in the picture vocabulary test of the WMLS.

Most children who responded to this item correctly, responded incorrectly to �ve or six items before and to all items after

‘‘llama’’ This item is located among items expected to be of

higher dif�culty for �rst-grade children; however, for many children in this study the llama was a familiar animal, since it is originally from the Andean region It is possible that because of its location in the test, there were children who lost the opportunity of responding to this item correctly, so they may have lost an additional point in their scores On the other hand,

if it was meant to represent higher dif�culty, this item appears not to be useful for this population

The lack of culture-speci�c language assessment instru-ments for culturally and linguistically diverse populations in developing countries and in the United States may have a negative and long-lasting in�uence on children for whom educational services may make the difference in their lives

Results obtained from translations of tests that have not been

standardised or adapted appropriately for the population in

which they will be used may lead to misclassi�cation and

incorrect placement or may exclude children from intervention

programmes Some efforts have been made (Atkin, 1989;

Landers & Kagitcibasi, 1990); however, there is still much

work to do to af�rm that language assessments for culturally

and linguistically diverse populations are free of reliability and

validity problems It is necessary to develop measures that

describe and explain the children’s current development and

functioning in their environments (Baine, 1990) Research will

be enhanced in Latin America when appropriate instruments

are employed

Conclusions

The importance of family processes in enhancing children’s

language and reading skills is supported by the results of this

study Parental expectations appeared to be strongly associated

with both children’s language and reading skills This

signi�cant association, in this population, may be interpreted

as a proxy for caregiving practices at home, other than those

measured in this study, that promote children’s development

and learning

Also, the study results are con�rmatory of �ndings in

numerous studies around the world about the bene�ts of early

childhood education on the developmen t and education of

children living in poverty In this study, children with no

preschool experience performed least well on letter and word

identi�cation However, it would be necessary to investigate

further why this signi�cant relationship was not found on

reading comprehension scores and on the oral language

measures, vocabulary, and verbal analogies Future research

will also be needed to examine in which ways type and quality

are related in preschool services offered to Peruvian children

Finally, with regard to research methodology, this study

presents evidence of the need for developing culturally

appropriate instruments to measure oral language and reading

in Peruvian Spanish-speaking populations The translated

version of the oral language test had limitations in evaluating

the speci�c characteristics of Peruvian children in this study

Furthermore, even the reading comprehension instrument

originally designed to be used in the Spanish-speaking

population in Chile needed to be adapted to the Peruvian

culture and Spanish usage in order to provide reliable

information Much work remains to be done in this area

Given the paucity of literature on South American and

speci�cally on Peruvian children, this study contributes to

expand the knowledge base and provides direction for future

research on this population

Manuscript received December 1998 Revised manuscript received June 2001

References

Alliende, F., Condemar´n, M., & Milicic, N (1993) Prueba de ComprensioÂnõ

Lectora de Complejidad Lingu¨istica Progresiva: Formas Paralelas(4ta Edicio´n).

Santiago: Ediciones Universidad Cato´lica de Chile.

Atkin, L (1989) Analysis of instruments used in Latin America to measure

psychosocial development in children from 0 to 6 years of age Paper prepared for

the Consultative Group on Early Childhood Care and Development New

York: UNICEF House.

Baine, D (1990) Guide to the development, evaluation, and/or adoption and

modi�cation of tests for early childhood education in developing countries In

M.J Thorburn & K Marfo (Eds.), Practical approaches to childhood disability in

Canada: Project SEREDEC/Jamaica: 3D Project.

Barber, B.L (1988) The in�uence of family demographics and parental

teaching practices on Peruvian children’s academic achievement Human

Development , 31, 370–377.

Boocock, S.S (1995) Early childhood programs in other nations: Goals and

outcomes The Future of Children: Long-term Outcomes of Early Childhood

Programs , 5, 94–114.

Bradley, R.H., Whiteside, L., Mundfrom, D.J., Casey, P.H., Kelleher, K.J., & Pope, S.K (1994) Early indications of resilience and their relation to experiences in the home environments of low birthweight, premature children

living in poverty Child Development, 65, 346–360.

Bryant, D., & Maxwell, K (1997) The effectivenes s of early intervention for

disadvantaged children In M.J Guralnick (Ed.), The effectiveness of early

Bryant, D.M., Peisner-Feinberg, E.S., & Clifford, R.M (1993) Evaluation of

University of North Carolina at Chapel Hill, Frank Porter Graham Child Development Center.

Burchinal, M.R., Campbell, F.A., Bryant, D.M., Wasik, B.H., & Ramey, C.T (1997) Early intervention and mediating processes in cognitive performance

of children of low-income African-American families Child Development, 68,

935–954.

Burchinal, M.R., Lee, M., & Ramey, C (1989) Type of day-care and preschool

intellectual development in disadvantaged children Child Development, 60,

128–137.

Burchinal, M.R., Roberts, J.E., Nabors, L.A., & Bryant, D.M (1996) Quality of

center child care and infant cognitive and language development Child

Development , 67, 606–620.

Burchinal, M.R., Roberts, J.E., Riggins, R., Zeisel, S.A., Neebe, E., & Bryant, D (2000) Relating quality of center-based child care to early cognitive and

language development longitudinally Child Development, 71, 339–357.

Campbell, F.A., & Ramey, C.T (1994) Effects of early intervention on intellectual and academic achievement: A follow-up study of children from

low-income families Child Development, 65, 684–698 [Special Issue: Children

and Poverty].

Carpenter, P.G., & Fleishman, J.A (1987) Linking intentions and behavior:

Australian students’ college plans and college attendance American

Educa-tional Research Journal , 24, 79–105.

Consortium of Longitudinal Studies (1983) As the twig is bent Lasting effects

of preschool programs Hillsdale, NJ: Lawrence Erlbaum.

DeBaryshe, B.D., Rodarmel, S.L., Daly, B.A., & Huntley, L (1992) Shared

picture book reading in the home: A language enrichment program for Head Start

GA.

Dickinson, D., & Snow, C (1987) Interrelationships among prereading and oral

language in kindergartners from two social classes Early Childhood Research

Quarterly , 2, 1–25.

Ferreiro, E., & Teberosky, A (1982) Literacy before schooling Exeter, NH:

Heinemann Educational Books.

Gill, S (1997) Educational expectation s and school achievemen t of children in

poverty Dissertation Abstracts International Section A: Humanities and Social

Sciences , 58 (5-A), 1579.

Goyette, K., & Xie, Y (1999) Educational expectations of Asian-American

youths: Determinants and ethnic differences Sociology of Education, 72,

22–36.

Halle, T.G., Kurtz-Costes, B., & Mahoney, J.L (1997) Family in�uences on

school acheivemen t in low-income, African American children Journal of

Educational Psychology , 89, 527–537.

Hauser-Cram, P., Pierson, D., Walker, D., & Tivnan, T (1991) Early education

in the public schools San Francisco, CA: Jossey-Bass.

Howes, C., Phillips, D.A., & Whitebook, M (1992) Thresholds of quality: Implications for the social development of children in center-based child care.

Child Development , 53, 449–460.

Landers, C., & Kagitcibasi, C (1990) Measuring the psychosocial development of

Italy: UNICEF International Child Development Centre.

Margolis, P.A., Greenberg, R.A., & Keyes, L.L (1992) Lower respiratory illness

in infants and low socioeconomic status American Journal of Public Health, 82,

1119–1126.

Martin, S.L., Ramey, C.T , & Ramey S.L (1990) T he prevention of intellectual impairment in children of impoverished families: Findings of a randomized

trial of educational daycare American Journal of Public Health, 80, 844–847.

Mau, W (1997) Parental in�uences on the high school students’ academic achievement: A comparison of Asian immigrants, Asian Americans, and

White Americans Psychology in the Schools, 34, 267–277.

Montenegro, H (1992) Retardo cultural El nin˜o con desventajas socio-culturales [Socio-cultural retardation The child with socio-cultural

disad-vantages] In G Alvarez (Ed.) Retardo socio-cultural Santiago, Chile:

Corporacio´n de Promocio´n Universitaria.