Recognizing the important contributions of school administrators to school performance, a nascent but growing literature in education policy seeks to identify the factors that differentiate effective from ineffective school principals (Branch, Hanushek, & Rivkin, 2012; Grissom & Loeb, 2011). A general finding from this research is that few observable characteristics of the kind typically available in administrative data sets are strong predictors of principal performance once school-level factors are taken into account (e.g., Clark, Martorell, & Rockoff, 2009). Researchers have thus begun to leverage alternative data sources, together with administrative records, to link otherwise unobservable aspects of principal skill or practice to student or teacher outcomes.

This study utilizes just such a unique data source to “get inside” a critical aspect of school leadership: principals’ time use within the school day. Over three different school years (2008, 2011, and 2012), we have sent trained observers into a stratified random sample of approximately 100 schools in Miami-Dade County Public Schools (M-DCPS), the nation’s fourth-largest school district, to shadow school administrators over full school days and record detailed information about principal time allocation. This information includes the specific tasks they engage in, who they interact with, what the content of those interactions are, and where they are located. In one of the years, we complement our shadowing of principals with shadowing of their assistant principals on the same day as well, which permits us to look closely at the distribution of leadership tasks and other aspects of the interactions among the school’s formal leadership team. Additionally, we combine the shadowing data with rich administrative data provided to us by the district—including information both on school personnel background and a variety of school outcomes—and with surveys administered to school administrators and teachers in each of the study years. The richness of these data allow us both to link time allocation characteristics to administrator performance and to contextualize their time use within the challenges of their various school environments. For example, we find, perhaps not surprisingly, that high school principals encounter a larger number of different tasks and sustain individual tasks for less time than their elementary and middle school colleagues.

Prior studies of principal time use have found that effective principals allocate more time to traditional management tasks, such as budgeting or maintaining facilities, rather than on the kinds of instructional tasks (e.g., teacher observations) that instructional leadership research has suggested to be most useful (Horng, Klasik, & Loeb, 2010). Previous studies, however, have been based on cross-sectional data. Preliminary analysis of our longitudinal data uncovers a more important role for instruction management in promoting school improvement, though some areas matter more than others. In particular, while principals in the high schools that improve most over time spend no more of their day on classroom observations or walk-throughs, they spend more time planning or managing staff professional development.