Study of Language and Literacy Development
Co-Principal Investigator: Catherine
E. Snow, Ph.D.
This page provides a description of the Home-School Study of Language and Literacy Development and a brief summary of selected findings:
The Home-School Study of Language and Literacy Development is a longitudinal research project that has been on-going for the last thirteen years. The original purpose of the Home-School Study of Language and Literacy Development was to investigate the social prerequisites to literacy success, identifiable in both home and school interactions, of a group of racially diverse, English-speaking children from low-income families growing up in the Boston area. A particular focus was on the relationships between early experiences with decontextualized language and later reading comprehension. A low-income sample (original sample = 83; present sample = 57) was selected for the study because of the assumption that such a sample was at elevated risk of experiencing difficulties with literacy; we also assumed that there would be variation in achievement across the group and that some of the students would be scoring at or above national norms by 4th grade. Both of these assumptions proved to be correct.
Research questions guiding us into the high school years include:
Is success recoverable? Does high school offer a student permission to do well academically when middle school did not, either because the social scene has changed or because there is a better fit between the student's needs and the high school's offerings?
Is success redefinable? Can high school provide avenues to success that middle school was unable to provide, for example in vocational courses or other courses that would tap a student's less traditional abilities?
This sample of 57 (going into high school) consists of 31 girls (54%) and 26 boys (46%) evenly split between Cohort 1 (49%) and Cohort 2 (51%). Twenty-one percent (21%) of the children are African-American, 67% are White, 5% are Hispanic, and 7% are bi-racial. Maternal education level at baseline ranged from 6 years of schooling to college enrollment. Thirty-five percent had not completed high school, 37% had received only a high school diploma, and 28% had attended some post-high school education. This sample was also distributed across the income levels represented by the original sample: at the first home visit 40% of these families reported an income of less than $10,000, 11% reported on income of $10-15,000, 19% reported an income of $15-20,000, and 21% reported an income of over $25,000. At baseline, thirty-nine percent of these families received AFDC as their main income source, 60% received income through employment (23% from primary caretaker, 37% from spouse), and one family reported child support as the main source of income.
Over the past thirteen years, the Home-School Study of Language and Literacy Development has collected data on a group of children (original sample = 83; present sample = 57) from age three through age fifteen, who are growing up in low-income families. These children live in families of differing configurations, with parents of varying educational achievements, in low-income or mixed neighborhoods. Over the course of the study they have experienced a variety of preschool (Head Start, private preschool, and day care programs), elementary, middle, and now high school (public, parochial, and private) learning environments. The project has been funded by the Ford Foundation, the Spencer Foundation, the Department of Health and Human Services (Head Start Grant), and the William T. Grant Foundation. From age three through seventh grade, we collected the following types of data for the study:
Interviews with mothers concerning life circumstances, child-rearing beliefs, and literacy-related activities throughout the study period;
Data on interactions between the children and their teachers and other children in their preschool classrooms, and data on their teachers' language use in their preschool classrooms;
Observational data focusing on language and literacy activities from the preschool and elementary school classrooms attended by the children;
Interviews with the children's teachers concerning the children's achievements, the teachers' educational practices, and, specifically, their literacy practices throughout the study period;
Tests of the children's language skills and school achievement, including literacy, from kindergarten through the seventh grade;
Interviews with the children's mothers and teachers related to affective and emotional factors in sixth and seventh grade;
Interviews with the students related to affective and emotional factors in sixth and seventh grade.
A primary focus of the earliest phase of the Home-School Study was on the home and preschool factors that would predict children's language and literacy skills in kindergarten. In a series of analyses, we were able to predict the children's scores on emergent literacy, vocabulary, and narrative production on the kindergarten SHELL (the School-Home Early Language and Literacy battery; see Snow et al. (1995)), from a variety of sources, including home interview information concerning home support for literacy, as well as some measures of decontextualized language and rare word use by mothers; classroom quality, extended talk, and vocabulary in preschool; and classroom quality in kindergarten. In combined regression analyses, science process talk, home support for literacy, and kindergarten classroom quality explain 35% of the variance in the scores on the narrative production task; home support for literacy and rare words in mealtimes and toy play explain 27% of the variance in emergent literacy; and home support for literacy, rare words at mealtimes and toy play, and kindergarten classroom quality explain 50% of the variance in the vocabulary measure (Dickinson, Snow, Roach., Smith, & Tabors, 1998).
The main conclusion from these analyses is that it is possible to find both home and school precursors for the children's literacy and language skills when they were five years old.
We continued to administer versions of the SHELL throughout the elementary school grades to track the children's achievement in language and literacy (see Publications, Snow (1995)). Three particular tasks were repeatedly administered: the Wide Range Achievement Test – Reading (WRAT) (Jastak, & Wilkinson, 1984), the Peabody Picture Vocabulary Test (PPVT) (Dunn, & Dunn, 1981), and a definitions task in which the students were asked to give definitions for ten common words. In 4th grade we administered the reading section of the California Achievement Test (CAT) (1992). Using growth modeling, we have found that student's individual growth trajectories, both intercept and slope, on word recognition (WRAT) (r2=.66), vocabulary (PPVT) (r2=.50), and decontextualized language (definitions) (r2=.29) were all significant independent predictors of scores on the 4th grade reading comprehension test (CAT). In combination, word recognition intercept and slope, and vocabulary intercept explained 74% of the variance in the 4th grade reading measure. (Dickinson, Tabors, & Roach, 1996).
In a further set of analyses, we are looking at the influence
of home and classroom factors from the preschool period on the students'
scores on the 4th grade reading comprehension test (CAT). As can be seen
in the table below, there are significant correlations between demographic
variables, home language and literacy variables, and preschool and kindergarten
language and quality variables with the 4th grade reading
*These variables include imputed values
When these variables are combined in a regression analysis, it is possible to explain 40.6% of the variation in the CAT scores based on income at first home visit, home support for literacy, rare word use in 4 year old preschool classrooms, and 4 year old preschool classroom quality.
A further, but still preliminary analysis, examined the relationship between the early home and school language and literacy variables and the growth trajectories in word recognition, vocabulary, and decontextualized language. Both the home and the school early language and literacy measures predict children's development in all three literacy domains. Home factors contribute more strongly, on average, than school factors. Of all the factors included in the analysis, the strongest predictor of children's literacy development was support for literacy in the home, which predicted average vocabulary (t=3.18, p<.01), average word recognition (t=2.54, p<.05), word recognition over time (t=3.59, p<.01) average decontextualized language (t=4.02, p<.001) and decontextualized language over time (t=2.82, p<.01) (Roach, 1999).
The intervention, Early Access to Success in Education (Project EASE), founded by Gail Jordan of the White Bear Lake School District in Minnesota, has proved to be successful in improving language skills for children at moderate risk. Project EASE was designed to increase the frequency and quality of language interactions through book-centered activities and to give parents information about and opportunities for engagement in their children’s developing literacy abilities. An article by Jordan, Snow, and Porche, "Project EASE: The Effect of a Family Literacy Project On Kindergarten Students’ Early Literacy Skills," describes the effects of this intervention (2000, Reading Research Quarterly, 35,4).
The Collaborative Language and Literacy Instruction Project (CLLIP), founded by Daniel Pallante of the Ohio Educational Development Center, is an extensive teacher training program grounded in theoretical and empirical work that emphasizes the primacy of language development as a critical foundation of early literacy and school success.