Stability of Teacher Behaviors and the Effects of Children’s Entering Skills

• In Event: Within Classroom Variation is not Noise: Predictors and Impacts of Varied ECE Experiences

Abstract

Early childhood classrooms are complex environments, consisting of students and teachers, all of whom have important differences among them. In order to capture the quality of an ECE environment, there are important issues to address. The first relates to teacher behavior: how stable are teacher behaviors across the school year? Does it matter in what settings behaviors are observed? Second, are teachers’ behaviors affected by qualities of the children in the classroom (e.g., entering academic and executive function skills)?
Generalizability Theory (Cronbach, Gleser, Nanda, & Rajarathnam, 1972) allows for the examination of multiple sources of variance simultaneously to understand the consistency and generalizability of measurements (i.e., reliability and validity), it allows for the simultaneous estimation of error based on an array of predefined sources of (e.g., occasion, setting) which is particularly important for applied, field-based research.
One goal of this paper is to apply Generalizability Theory to quantify potential sources of measurement error/variance in examining the quality of 80 public prekindergarten classrooms predominantly serving economically disadvantaged families. A snapshot system to capture children’s and teachers’ behaviors was used across the entire school day and across all children and adults in the classrooms. The impacts of occasion (fall, winter, vs. spring) and setting (teacher-directed whole and small groups vs. child-directed centers) on the amount and quality of teacher instruction were examined. Additionally, relations between the classroom’s initial scores on an array of academic achievement and executive function assessments, as well as variance in scores within a classroom, and teachers’ quality and quantity of instruction were examined.
Analyses of variance associated with differences among teachers, by occasion, and by setting (Table 1) indicate that for quantity of instruction the largest source of variability was between settings (26.3% in teacher-directed vs. 12.3% child-directed), including individual variability between teachers in how much they allocated time to teacher-directed and child-directed instruction. Teachers remained consistent across occasions in how much they instructed. For the quality of instruction, individual teachers varied in the quality of their instruction across the year and across settings. Thus, valid measures of classroom instructional quantity need to consider sampling across multiple settings while measures of quality must sample across both occasions and settings. Examination of variability in quantity and quality of instruction attributed to differences in children’s entering skills (Figure 1) indicated that higher entering academic skills, were associated with more teacher instruction, but not the quality of the instruction, while higher entering executive function skills were positively related to quality, but not quantity. Additionally, greater variability in skills among children within a classroom was related to lower quality instruction for both entering academic and executive function skills. In other words, the more dissimilar children were from each other at the onset of the year, the lower the quality of instruction provided.
For the presentation, implications of these findings for early childhood education practice and policy will be discussed.

Authors

• Dale C Farran, Vanderbilt University
  Presenting Author

• Kimberly Turner Nesbitt, University of New Hampshire
  Non-Presenting Author