Project Activities
The researchers used secondary data to explore three research questions:
- What subgroups of DLLs, based on levels of cognitive, oral language, literacy, and math achievement in both English and Spanish, are formed when examining the heterogeneity of Latino DLLs attending Florida Head Start programs?
- What subgroups of DLLs are formed, based on Latino DLLs’ initial levels and rates of achievement growth as they transition from Head Start into kindergarten? Which malleable home and classroom language factors and classroom quality dimensions are associated with trajectories?
- What subgroups of DLLs are formed, based on initial levels and rates of growth for social emotional skills?
To address these questions, the researchers first prepared the dataset and constructed two cohorts (younger students who had 2 years of Head Start and older students who had 1 year of Head Start and 1 year of kindergarten) and then conducted analyses to address the research questions.
Structured Abstract
Setting
The data came from Head Start and kindergarten classrooms serving Latino DLLs in five large urban districts in the south and central regions of Florida.
Sample
The sample included 396 Latino DLLs, across two cohorts, enrolled in Florida Head Start programs from 2009-2011. Cohort 1 attended Head Start and kindergarten. Cohort 2 attended Head Start for the 2 years.
Factors related to learner outcomes included maternal education, language use at home, language use in the classroom, teacher education, and teacher hours of training on working with DLLs.
Research design and methods
The data used for this secondary data analysis included 396 Latino DLLs who were enrolled in Head Start programs in Florida between 2009-2011. The researchers created cohorts from these data and employed a cross-sectional longitudinal design. These cohorts included a younger cohort (n = 128) of 3-year-old prekindergarten (preK) through the spring of 4-year-old preK students and an older cohort (n = 258) of 4-year-old preK through the kindergarten students. The researchers used assessment data from three time points for each student cohort. Using latent profile analysis, the researchers identified profiles of 320 DLLs at the end of the 4-year-old preK year. They used growth mixture models to identify growth profiles for the older cohort through the spring of kindergarten, as well as to identify growth profiles of social-emotional development as the older cohort transitions from the spring of preK through the spring of kindergarten.
Key measures
The key cognitive, linguistic, and academic child-level outcomes were scores from the following subtests of the Woodcock-Johnson III/ Batería III Woodcock-Muñoz administered in both English and Spanish: picture vocabulary, oral comprehension, letter-word identification, spelling, applied problems, quantitative concepts, spatial relations, visual matching, and picture recognition. The key social-emotional child-level outcomes were scores from the Penn Interactive Peer Play Scale (PIPPS) and the Preschool Learning Behavior Scale (PLBS). Malleable classroom factors included CLASS scores and responses from a teacher demographic survey. Home factors were obtained from a parent demographic survey.
Data analytic strategy
The researchers accounted for the nesting of DLLs within classrooms in all statistical models (i.e., multilevel modeling). They used latent profile analysis to examine cross-linguistic school readiness profiles and used growth mixture modeling to examine profiles of latent growth trajectories for language, literacy, math, and social emotional development. After deciding on the number of profiles, the researchers included home and classroom covariates in that model.
Key outcomes
The main findings of this project are as follows:
- Latino dual language learners (DLLs) enrolled in Head Start programs in the southeastern United States represent four different profiles of learners including English dominant, Spanish dominant, balanced average, and emerging bilinguals. Balanced average children outperform their peers on child outcomes in both English and Spanish. Children in the balanced average and Spanish dominant profiles are typically in classrooms with teachers who report higher levels of education and more training hours related to working with DLLs (Lopez et al., 2024).
People and institutions involved
IES program contact(s)
Project contributors
Products and publications
Study registration:
Publications:
ERIC Citations: Find available citations in ERIC for this award here.
Select Publications:
López, L.M., Foster, M.E, Sutter, S., Nylund-Gibson, K., & Arch, D.N.A. (2024). Subgroups Within a Heterogeneous Population: Considering Contextual Factors that Influence the Formation of DLL Profiles in Head Start. Journal of Educational Psychology, 116(1), 123-138. DOI: 10.1037/edu0000825
Available data:
Data will be available publicly through a data sharing agreement facilitated through LDBase (https://ldbase.org/) once manuscripts for the main research questions have been published in a peer-review journal. Until then data can be requested by emailing the PI at lmlopez@usf.edu.
Questions about this project?
To answer additional questions about this project or provide feedback, please contact the program officer.
