Library space shortage has been a pressing problem for several decades
. A variety of alternatives, such as remote storage and weeding, have
been proposed and attempted. Previous research in this area has been f
ragmented. Some research is concentrated on criteria for identifying r
arely used material for weeding; in other research, costs of alternati
ves to the traditional library are compared. In this paper a new model
that takes a more complete approach to the problem is proposed. We in
corporate the prediction of future demand and cost analysis into one m
odel. The model includes random effects that cause errors in the selec
tion process. By applying data of a hypothetical collection, we show h
ow this model helps the collection development librarian determine the
most economic approach.