Identifying Assessment Approaches for Dual Language Learners: Evidence From a National Study of Head Start

• In Event: Data for Decisions: Family and Child Experiences Survey as a Model of Policy-Driven and Policy-Guiding Research

Abstract

Meeting the needs of children who are dual language learners (DLLs) and their families is a top priority for the Office of Head Start (OHS). Thus, we undertook a partnership with the Administration for Children and Families to ensure the Head Start Family and Child Experiences Survey (FACES) would capture the skills of DLLs in a manner that best met Head Start’s information needs. In this paper, we will describe (1) the process of partnering with OHS and measurement experts to identify analyses for determining whether assessment procedures could be improved and (2) results of those analyses.

The Partnership Process. We first met with representatives of OHS to identify Head Start program standards related to DLLs. For example, programs must assess DLLs in a manner that allows them to demonstrate their knowledge and skills regardless of home language. Subsequently, an expert panel recommended a series of analyses to determine if FACES language routing procedures also allowed for meeting this standard. FACES uses the Simon Says (receptive language) and Art Show (expressive language) subtests from the preLAS 2000 as an English screener. Children who make five consecutive errors on both 10-item subtests are routed through the assessment along a Spanish or non-English path depending on home language. Language paths for each child can change across waves of data collection as English skills improve. Regardless of screener performance, all children complete the English One-Word Picture Vocabulary Test (EOWPVT), a conceptually scored assessment of vocabulary.

Analyses to Identify the Optimal Assessment Approach. We are conducting analyses to answer the following research questions:

• What are the relationships between performance on the preLAS and scores on other assessment measures?
• What are the relationships between performance on the screener, children’s home language, and the likelihood that children respond to the EOWPVT in English or Spanish?
• Can using information about language of response on the EOWPVT improve routing?
Preliminary Results. Table 1 shows that children with an English home language have the highest total preLAS score, followed by DLL children who passed the screener in fall 2009 (14.8 and 10.0, respectively). DLL children who did not pass the screener in either fall 2009 or spring 2010 had scores of 2.4 points or less at both time points. Table 2 shows the performance of children on the cognitive assessments in fall 2009 and spring 2010 by language of assessment. First, children assessed in English at both waves outperformed children who switched to English in the spring on all spring assessments (with the exception of the TVIP). Second, in the fall, children assessed in Spanish during both waves performed similar to or slightly lower than children who switched from a Spanish to an English assessment in the spring. Third, children assessed in Spanish at both times performed more poorly in the spring than either the children assessed in English at both times or children who switched from a Spanish to English assessment by the spring.

Author

• Nikki Aikens, Mathematica Policy Research