How postings information is used for inverted file searching was inves
tigated by comparing searches, made by postgraduate students at the De
partment of Information Studies, of the LISA (Library and Information
Science Abstracts) database on CD-ROM with and without postings inform
ation. Performance (the number of relevant references, precision and r
ecall) was not significantly different but searches with postings info
rmation took more time, and more sets were viewed, than in searches wi
thout postings. Postings information was used to make decisions to nar
row or broaden the search; to view or print the references. The same t
echniques were used to amend searches whether or not postings informat
ion was available. Users decided that a search was satisfactory on the
basis of the search results, and consequently many searches done with
out postings were still considered satisfactory. However, searchers th
ought that the lack of postings information had affected 90% of their
searches. Differences in search performance and searching behaviour we
re found in participants who were shown to have different learning sty
les using the Witkin's Embedded Figures Test and the Lancaster Short I
nventory of Approaches to Learning Test. These differences were, in pa
rt, explained by the differences in behaviour indicated by their learn
ing styles.