Danny P. Wallace

The relationship between journal productivity and obsolescence

This study examined the relationship between journal productivity and journal obsolescence for a database of references from articles dealing with desalination. Although these two variables have often been studied in isolation, no previous studies have examined their interaction within a single subject literature. It was hypothesized that those journals that were most productive would, on the average, have relatively short active lives, and that as journal productivity decreased, the average active lives of the articles contributed by a journal would increase. The number of references to a particular journal in the database was used as a measure of that journals productivity. The measure of obsolescence used was the median age of the references to a particular journal. The hypothesized inverse linear relationship was not found to hold, although the data did exhibit an inverse tendency. It was found that highly productive journals did tend to have low journal median citation ages, and that high journal median citation ages were always associated with journals that were unproductive in terms of the numbers of references to those journals in the database. These extreme cases appeared to be distributed in a hyperbolic manner. The remaining journals, which were not highly productive and did not have high journal median citation ages, appeared to be distributed in a random manner.

The Use of Statistical Methods in Library and Information Science.

This study compared the use of statistics in 99 journals from four subject areas: library and information science, education, social work, and business. It was found that journals in library and information science produced more articles making no use of statistics than did journals in the other three subject areas, and that only in library and information science were there more articles using descriptive techniques only than articles using inferential techniques. A comparison of the mean number of articles per journal using no statistics, descriptive statistics only, and inferential statistics indicated that the mean number of articles per journal using inferential statistics was much lower for library and information science than for the other subject areas. The only inferential technique not used significantly less in library and information science than in the other subject areas was correlation, one of the simplest of inferential techniques.

Measurement of subject scatter in the superintendent of documents classification

Abstract The hypothesis that the amount of dispersion of documents occurring when a collection is reclassified from the Superintendent of Documents Classification (SuDoc) to the Library of Congress Classification (LCC) is insignificant is not supported. This suggests that the SuDoc scheme is inappropriate for topical questions. The rank order statistics show no relationship between the document orderings produced by the two schemes, and in no case did an examined SuDoc class contain less than 48 percent of the LCC main classes. Library of Congress MARC records containing SuDocs numbers for the Department of the Army, the Department of Agriculture, the Department of State, the Library of Congress, and a broad general sample were sorted by both schemes and the resulting orders compared. A new measure, normalized average minimum document movement was proposed and computed, as were existing measures of dispersion resulting from reclassification.