This paper summarizes the findings of a recent study on the indexing p
ractices used in the Library and Information Science Abstracts (LISA)
database. The indexing terms (DE), the date each record was added to t
he file (DA), the Accession Number (AN), and the Classification Code (
CC) of each record were extracted from the complete CD-ROM database. A
djustments to standardize the DE terms were made, and the adjusted dat
a set was analyzed for average number of headings per record per DA ye
ar, the rank frequency and rank size distribution of DE classes, and t
he frequency distribution of the number of DEs per record per DA. The
results show that a large number of headings are used once or twice ov
er the whole database. The years in which DE terms first appeared was
analyzed. A comparison of the use of a particular Classification Code
with the DE terms in each record was also undertaken. There was a stro
ng but not total match between the CCs and DEs used. Some attention is
given to the chain indexing procedure used by LISA to account for the
pattern of DE usage. The concluding section looks at scope for furthe
r research on LISA and other databases.