Dissertation | Abstract > Introduction

Auditing catalogue quality by random sampling


A method is proposed for auditing the quality of cataloguing, developing work by Ann Chapman of UKOLN towards a simple tool for libraries to assess the accuracy of their bibliographic records. It is intended to produce a quantitative measure which can be used as a performance indicator.

The approach taken is to compare a random sample of catalogue records against the items they describe, recording discrepancies on the worksheet provided. A second sample is generated by taking items a fixed distance to the left of those in the first sample and these are searched for on the catalogue to check for uncatalogued items and duplicate records.

The technique is influenced by a review of the literature on quality control in bibliographic databases and library catalogues. Brief instructions are followed by a detailed examination of implementing the audit. In particular, there is a discussion of what makes a valid sample and effective substitutes for true random sampling. A pilot in the library of the University of Bath demonstrated the practical difficulties in generating a random sample. The sample obtained was not strictly random but its findings were confirmed by a smaller systematic sample.

The results of the pilot highlight the importance of profiling a library's collection prior to an audit to allow for local idiosyncrasies. The technique is shown to be effective for picking up errors in descriptive cataloguing but less sound for name and subject headings and a brief survey of typical errors in these fields is given.

A simplified technique is proposed, reducing the amount of detail to be recorded and removing the test for uncatalogued and duplicate items. It is recommended that the audit be considered just one tool in a toolkit for evaluating catalogue quality.


I would like to thank Ann O'Brien, my supervisor at Loughborough, for her guidance and encouragement; Ann Chapman of the Bibliographic Management group at UKOLN for the idea and for arranging for the pilot study to take place in Bath; and the staff of the University of Bath's Library and Learning Centre, especially Ann Cox, systems librarian, who produced the first sample, Diana Gibney, technical support librarian, and Gavin Rea, deputy librarian.

I am also grateful for the support of my parents and of my partner, Alice Dryden.

Dissertation | Abstract > Introduction

Owen Massey McKnight <owen.mcknight@worc.ox.ac.uk>