Using classroom observations to evaluate science teaching: Implications of lesson sampling for measuring science teaching effectiveness across lesson types

Abstract

Despite the prevalent use of observational measures in teacher evaluation systems, research has only recently begun to take into account how aspects of the instructional environment and lesson sampling design may interact with teachers’ scores on these measures. Instead, one assumption guiding current evaluation systems is that the variation in teachers’ instructional practices captured by the lessons sampled is a valid measure of their teaching quality. Our study investigates this assumption by examining whether scores from observational measures of science teachers’ instructional practices vary depending on lesson type. Results show that science teachers’ scores on certain observational measures vary across lab and nonlab lessons, suggesting that the claims one can make about science teaching effectiveness are dependent on the particular types of lessons observed. Given the increasing prominence and use of observational measures within evaluation systems, it is imperative to ensure that the observed lessons adequately represent variation in a teacher’s practice and are not subject to undue instability from the types of lessons included in the sample, especially in science where there are clear distinctions in the instructional practices teachers employ across lab and nonlab lessons. Implications for measuring science teaching quality are discussed.