This blog provides an overview of the main findings of a recently published article in Child Development, “Teacher-Child Interactions in Chile and their Associations with Prekindergarten Outcomes” by Diana Leyva, Christina Weiland, Clara Barata, Hirokazu Yoshikawa, Catherine Snow, Ernesto Trevino, and Andrea Rolla.
To receive a copy of the full article, email Diana Leyva.
Chilean preschool classrooms look very different from U.S. preschool classrooms. The child-staff ratio in Chilean public preschool classrooms is relatively high compared to the U.S. On average, there are two adults (a teacher and a teacher aide) for every 35 preschool children. In addition, small group activities such as activity centers are rare in Chilean preschool classrooms. Teachers promote children’s skill development mainly through whole-group activities such as free-drawing, free-play, and oral routines (singing and talking). Furthermore, classrooms tend to be small in size and typically include small desks and chairs for every child, all facing the teacher and the blackboard. Overall, classrooms are not thought of as places for children to move around.
These differences led us to wonder whether the current gold-standard for measuring quality of early childhood education in the U.S. – the CLASS- is valid in Chilean preschool classrooms. The CLASS (Classroom Assessment Scoring System; Pianta et al., 2008), is a national monitoring tool currently used by the federal Head Start program serving over 900,000 preschool-aged children from low-income families in the U.S. In addition, the CLASS has been integrated into the quality rating and improvement systems of several states in the U.S.
Validating this national monitoring tool is important because it helps us determine whether the data collected through it hold across populations with different sociocultural conditions. In this post, we summarize the results of the study we conducted, discuss some policy implications, and outline some next steps in this line of research.
We examined two types of validity in this tool: construct and predictive validity. The first type of validity deals with whether the data collected with this tool reflect the conceptual framework (the same characteristics) on which it is based. There are reasons to suspect it might not hold. For example, given the high staff-child ratio in Chile, it is possible that “good” instruction does not necessarily entail individualized attention and support for the child’s autonomy, two key characteristics of “good” instruction in the U.S. The second type of validity we measured deals with whether data collecting with this tool are predictive of gains in child outcomes. In other words, did Chilean children in classrooms rated as higher quality on the CLASS show more growth in their language, literacy, math, and executive function skills while enrolled in preschool?
We collected data on 91 public prekindergarten classrooms in 64 schools including over 1,800 children in Santiago, Chile. Classrooms were videotaped for the entire duration of a randomly chosen school day. We selected four 20-min video segments from each classroom to reflect the different times of the school day and activities (circle time, snack time, etc.). We coded these video segments using the CLASS, a 7-point scale that rates teacher-child interactions along 3 different dimensions: emotional support, classroom organization, and instructional support. A rigorous training and reliability process ensured that the same assessment criteria were used across video segments. We measured children’s language, literacy, and math using a standardized battery of tests. We also measured children’s executive function skills (ability to control and inhibit responses) using a number of tasks.
Two important results emerged from this study. First, we found evidence supporting the presence of the 3 dimensions of teacher-child interactions in the Chile data: emotional support, classroom organization and instructional support. Second, we found that while instructional support predicted children’s literacy and executive function, classroom organization predicted children’s language, literacy and math skills.
This is one of the first studies conducted outside the U.S. validating the CLASS. Validating this tool in a country like Chile is crucial for policy makers for three reasons:
- The ability to make informed evaluations of early childhood programs and investments is contingent upon having a common scale derived from cross-study assessments.
- Validating the CLASS helps understand how the quality of teacher-child interactions varies across settings and countries.
- Validating the CLASS can help in targeting early childhood education reforms and resources accordingly.
In Latin America, the CLASS is currently being used in studies of prekindergarten quality in countries such as Colombia, Mexico, Ecuador, Costa Rica, and Chile. Other international studies using the CLASS are being conducted in Norway, Finland, England, Spain, Germany, and Australia. We encourage researchers to examine whether and how the CLASS is a valid tool to assess quality of early childhood education in their countries before policy makers use its data to guide changes in policies, and investments in family programs and teacher professional development programs.
Finally, we should mention that the CLASS can be used, not only as an assessment tool to inform decisions about program quality, but also as a professional development tool to guide improvements in teaching practices in the classroom. Although most of the international studies currently underway use the CLASS as an assessment, it would be interesting to test whether and how this tool can be used to improve teacher practices in countries other than the U.S.
Who else is using CLASS in a low-/middle-income country context? We would love to exchange experiences and foster cross-fertilization of this important work. It is not impossible that the CLASS could fail to adequately capture what it is intended to. We, as a policy oriented research community need to know when and where this might be the case.