The information needs interview: a long way from library-use statistics

Many methods are used in the determining of information needs but most of them – and certainly the most commonly-used ones, are at best inappropriate and at worst misleading. Probably in no other area of information work does the research method influence so greatly what is discovered. It is generally recognised that information use data is no substitute for information needs data, yet researchers proceed to collect such data and continue to make bogus claims as to its provenance. The principal reason for this is that use data is easy to collect – often it's to hand (automatically produced by library/database management systems). And we must not overlook information managers' love affair with statistics either. Information needs data by its very nature can be obtained only through open-ended interviews and, less satisfactorily, through diaries and observation. The paper examines the structure of the information needs interview and makes some telling comparisons with use data – especially that gleaned from transactional log analysis.