A TIME-SERIES MODEL FOR ACADEMIC-LIBRARY DATA USING INTERVENTION ANALYSIS

Library time-series, as described in the literature, have been based o n monthly data that mask the highly variable nature of library circula tion and other internal statistics. This article presents a model of a n 8.4-year time-series of weekly academic library journal pickup data. Using MINITAB and SPSS procedures, various models, which included sea sonality and intervention variables, were tested. Although periodicity was observed by spectral analysis, it was not a factor in providing a better model. Rather, the best fit was achieved with a model which in cluded an intervention variable based on library hours.