A Dynamical View of High School Attendance: An Assessment of Short-term and Long-term Dependencies in Five Urban Schools

摘要

While school attendance is a critical mediator to academic achievement, its time dependent characteristics are rarely investigated. To remedy situation, this paper reports on the analysis of daily attendance rates in five urban high schools over a seven-year period. Traditional time series analyses were conducted to estimate short-range and cyclical dependencies in the data. An Autoregressive Fractional Integrated Moving Average (ARFIMA) approach was used to address long-range correlational patterns, and detect signs of self-organized criticality. The analysis reveals a strong cyclical pattern (weekly) in all five schools, and evidence for self-organized criticality in one of the five. These findings illustrate the insufficiency of traditional statistical summary measures to characterize the distribution of daily attendance, and they suggest that daily attendance is not necessarily the stable and predictable feature of school effectiveness it is conventionally assumed to be. While educational practitioners can probably attest to the many of the irregularities in attendance patterns as well as some of their sources, a systematic description of these temporal aspects needs to be included in our assessment of daily attendance behavior to inform policy decisions, if only to better align formal research in this area with existing local knowledge about those patterns.