Using multiple hypotheses in collection condition surveys

Abstract

This article assesses various problems associated with standardised methodologies and forms used in collection condition surveys. Three connected problems were explored: standardised methodology, diagnosticity and induction. These problems can result in interpretational bias when analysing condition data, and the possibility that data could be used to support contradictory conclusions. Using multiple working hypotheses is recommended as a means of relating assessment to specific collections: the process of hypothesis testing is one means of reducing the interpretational bias. The use of a set of hypotheses means that different explanations of the problems can be reviewed at the same time, adding focus to the survey. Generating and testing different hypotheses provides validity to conclusions and a rationale for collection management decisions.

Notes

Author for correspondence.