Quentin L. Burrell ¹

Affiliations
1 Statistical Laboratory, Department of Mathematics, The University of Manchester, Manchester M13 9PL, GB

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.

Full Text

     

Quentin L. Burrell ¹

Affiliations
1 Statistical Laboratory, Department of Mathematics, The University of Manchester, Manchester M 13 9PL, GB

Abstract

The theory of mixtures of point processes is used to provide models for library circulation data. The extent to which such models are able to mimic the observed correlation structure is investigated making use of the so-called Poisson regression ratio. Particular reference is made to the University of Saskatchewan database analyzed in Part 1 of this study.

Full Text