Files in this item
|(no description provided)|
|Title:||Qualitative Attributes and the Bradford Distribution (Bibliometrics, Evaluation)|
|Department / Program:||Library Science|
|Degree Granting Institution:||University of Illinois at Urbana-Champaign|
|Abstract:||Two related questions were the motivation behind this study: (1) Is it possible to find articles of similar quality in journals drawn from different zones of the Bradford distribution? (2) Are there any bibliometric indicators that may correlate well with the qualitative attributes of an article?
The subject selected was that of methanogenic bacteria. It is a fairly recent area of study whose literature has grown rapidly since 1980. The data base was assembled from bibliographic records retrieved online from BIOSIS, CA Search and MEDLINE. The search covered the years 1974-1978. Once irrelevant and duplicative items had been removed from the data base, it totaled 1212 periodical articles.
Assessment data were collected from judges via a questionnaire attached to the 189 articles, randomly selected from the data base. Articles were assigned at random to eight judges, a group of researchers in the field of methanogenic bacteria at the University of Illinois at Urbana-Champaign.
The Bradford distribution was used to group data into zones to allow testing of the null hypotheses.
The results of statistical tests on the data derived from the quality assessments of the experts support the hypothesis that: As judged by subject specialists, there is no significant difference in the importance of papers representative of the several zones of a Bradford distribution. Analysis of data on citation rates for articles in the sample do not support the hypothesis that: There are no significant differences among the various Bradford zones in terms of the extent to which papers representative of each zone are cited. Additional analyses yielded the following findings: Analyses of correlation between rating, citation, number of authors and number correlation between qualitative and quantitative attributes is in general below the moderate level (.6). The strongest correlations for zonal data were those between rating and citation data. The fourth zone showed the strongest correlation. A comparison of the first and the fourth zones revealed the following differences: (a) Zone 1 was characterized by average-to-high rating and average citation. (b) Zone 4 was characterized by low citation and high rating.
Based on frequency data, joint probabilities of occurrence of a selected event were establishd, limitations of the study are explained so are the areas for further study.
Thesis (Ph.D.)--University of Illinois at Urbana-Champaign, 1984.
|Date Available in IDEALS:||2014-12-15|