ISSN:
1588-2861
Source:
Springer Online Journal Archives 1860-2000
Topics:
Information Science and Librarianship
,
Nature of Science, Research, Systems of Higher Education, Museum Science
Notes:
Abstract Since the Groos droop of Bradford curves was reported, there has been a controversial explanation of its cause, i.e., that it is caused by an incomplete data set. In this study, a computer simulation was conducted to study the phenomenon. Incompleteness was characterized by two kinds of sampling, weighted and unweighted. Weighted sampling was used to simulate incompleteness of low productivity journals, unweighted sampling incompleteness at all productivity levels. Based on the result of 400 runs (two sampling methods x four sample sizes x ten data sets x five random runs), the hypothesis that the Groos droop is caused by incomplete data sets was rejected. The relationships between sample size, sampling method and the degree of the droop are also reported.
Type of Medium:
Electronic Resource
URL:
http://dx.doi.org/10.1007/BF02095349
Permalink