Poster #77 - Factors Explaining Differences in Classroom Quality: Evidence From Head Start

• In Event: Poster Session 03

Abstract

In the United States, access to high-quality early childhood education (ECE) is uneven and often linked to the socioeconomic backgrounds of children and families (Aguiar & Aguiar, 2020; Hatfield et al., 2015; McCoy et al., 2015). Given the increasing diversity of children participating in ECE settings (e.g., Office of Head Start 2018), it is important to understand whether certain groups of children are in classrooms with different observed classroom quality, the systemic factors that explain whether children are in high-quality classrooms, and the structural characteristics that shape equitable access to high-quality ECE. The current study aims to begin to fill these gaps in the field.

Drawing on Head Start Family and Child Experiences Survey (FACES) 2014 data, we explore possible explanations for why certain children (e.g., children of color, dual language learners) are in classrooms rated as lower quality compared with their counterparts and why certain classrooms (e.g., children with teachers of color) are similarly of lower quality.
First, we examined how observed classroom quality, as measured by the Classrooms Assessment Scoring System (CLASS; Pianta et al. 2008) and the Early Childhood Environmental Rating Scale–Revised (ECERS–R) Harms et al., 2005), differ by the racial/ethnic background of children (N=1,736) and teachers (N=597) and for classrooms with instruction in both English and another language (Table 1). We found that Black children and teachers were in classrooms rated as being of lower quality compared with White and Hispanic/Latinx children. Dual language learners were in classrooms rated as being of higher quality compared with English-only speaking children.

Next, we plan to examine whether differences in: (1) classroom composition (percentage of children from different racial/ethnic groups, with an IEP, with low baseline skills, in poverty, who are dual language learners); (2) structural measures of quality (class size, years of teacher experience, teacher education, turnover); (3) approaches to teaching (curriculum used; whether instruction is whole-group, small-group, individualized; frequency of literacy or math activities); (4) teacher characteristics (depression, job satisfaction); (5) supports from centers and programs (professional development) explain differences in classroom quality. We will compare mean values and other descriptive statistics among the groups we are comparing for each explanatory variable, focusing primarily on those for which we see a significant difference in subsequent analyses.

Once we determine the groups that are significantly different from each other in descriptive analyses, we will implement an Oaxaca-Blinder decomposition. This technique allows us to decompose average group differences into a component that is due to observed differences and an unexplained component (Jann, 2008). This type of analysis will tell us the magnitude of the differences in quality that would remain even if both groups had exactly the same characteristics. This analysis can feasibly be completed before the conference.

Results will contribute to our understanding of drivers for differences (including observer bias) in preschool classroom quality across groups. We will discuss implications of these results in the context of steps that programs and policymakers should consider to equitably promote high-quality preschool for all children.

Authors

• Tutrang Nguyen, Mathematica Policy Research - District of Columbia
  Presenting Author

• Christopher Hilton Jones, Mathematica Policy Research - District of Columbia

• Jessica Harding, Mathematica Policy Research - New Jersey

• Nikki Aikens, Mathematica Policy Research - District of Columbia