Browse By Day
Browse By Time
Browse By Panel
Browse By Session Type
Browse By Topic Area
High quality early learning experiences are critical for young children living in poverty (Dearing, McCartney, & Taylor, 2009). However, there is national concern that classroom quality is low in low-income toddler classrooms (Ruzek, Burchinal, Farkas, & Duncan, 2014). National policy initiatives, including Early Head Start (EHS) and Quality Rating and Improvement Systems (QRIS), have recently responded to this national need for access to quality for children from low-income households.
To improve quality, assessments of process quality such as the Classroom Assessment Scoring System-Toddler (CLASS-T; La Paro, Hamre, & Pianta, 2012) are used within EHS and QRIS. In addition, cut-points on the CLASS-T contribute to QRIS ratings of programs rated as high or low quality that have major policy implications (Schaack, Tarrant, Boller, & Tout, 2012). Despite widespread use of the CLASS-T, few studies have examined the construct validity of the CLASS-T for use in EHS and subsidized child care programs serving low-income, ethnically and linguistically diverse families. In addition, no studies to date, have empirically validated cut-points that QRIS programs use in practice.
The proposed project will address these gaps by first validating the CLASS-T in a sample of diverse toddlers attending EHS and subsidized child care programs. Then, associations between CLASS-T domains and child outcomes will be examined. Finally, the cut-points that are currently in use as part of the local QRIS program will be examined empirically to determine whether the cut-points used in practice identify thresholds of classroom quality.
Participants in the sample include 103 teachers with 824 children across both EHS and subsidized child care toddler programs participating in QRIS. Children in the sample included 52% boys between the ages of 18 and 36 months (M = 21.01, SD = 7.34) with 60% of the sample reported as Hispanic. The majority of teachers in the sample were female (96.9%) and Hispanic (80%). The measures examined in the study are included in Table 1.
Confirmatory factor analyses were examined to determine construct validity of the CLASS-T and to examine whether a 1-factor or 2-factor model would fit the data best. Both the 1-factor model, where all dimensions of the CLASS-T loaded onto one factor of teacher-child interaction quality, and the 2-factor model, in which there were two domains of teacher-child interaction quality, Emotional and Behavioral Support and Engaged Support for Learning, fit the data well. See Table 2 for model fit. To examine which factor structure fit the data best, we used the Satorra Bentler χ2 difference test. The results were significant, indicating that the 2-factor model fit our data best (χ2diff(1) = 27.06, p < 0.001).
In order to examine the relationship between the CLASS-T and child language, social-emotional, and cognitive development, multilevel regression analyses will be used. Additionally, piecewise regression multilevel analyses and b-spline models will be used to examine whether there is a threshold in the relationship between classroom quality and child outcomes in this sample. The child level data is currently being analyzed and those findings will be presented during the poster session.