Measurement and Modeling Choices in Studies of Time and Learning

Abstract

The consequences of using alternative measures of time and achievement are examined in an observational study of 18 elementary school classes. Of the four measures of time used—scheduled time, actual instructional time, engaged time and engaged rate—the engagement measures produced the more consequential effects of time for learning. Nonetheless, measures of allocated time are still important because they provide the constraints within which the results for engagement time must be interpreted. Mathematics achievement, measured as standardized and as chapter specific tests, was not consistently related to the time measures. This paper examines the extent to which the choice of the same model linking time and learning for all students was implicated in these inconsistent results.