Correlation Structure of Library Circulation Data:A Case Study:Part 1-The Empirical View

Abstract

When one looks at library circulation data collected over a two-year period and considers the regression of the mean number of loans in the second year on the observed number in the first, it has often been noted that the resulting graph appears to be approximately linear. This phenomenon has therefore become an established feature of models of library circulation. Recent work based on data collected over an eleven-year period at the University of Saskatchewan has questioned whether the perceived linearity is not more illusory than actual. This study reconsiders the Saskatchewan data using alternative simple regression analyses to more clearly reveal the regression structure.