Dissertation Information for Theodore Shih-Shu HuangNAME:
DEGREE:
DISCIPLINE:
SCHOOL: ADVISORS: COMMITTEE MEMBERS: MPACT Status: Fully Complete Title: EFFICACY OF CITATION INDEXING Abstract: Three topics of current research interest were submitted by three geneticists, one each. A backward search by checking documents was done for one topic. A forward search in a citation index and a search in subject indexes were done for all three topics. To compare the search results for efficacy, relative recall ratio, relative rejection ratio, noise ratio, redundancy ratio, unit search decision, and unit search time were used as evaluative measures, besides the one basic measure, pertinence ratio. For the pertinence judgment of references retrieved in the backward search, citing author’s testimony was used. For the pertinence judgment of references retrieved by searching the citation index and the subject indexes, the retrieved documents were read in entirety by the scientist who submitted the search topic. Pertinence was defined as “the presence in the document of information that is within the scope of the topic and is also useful to research on the topic.” The scope of each topic was defined by the scientist prior to the searches. Two hypotheses were formulated before the data were collected. The backward search by checking documents affirms the first hypothesis that a backward search, without the elimination of noise at each step, will retrieve an overwhelming number of references, most of them being noise, and with the noise increasing at each step. The results of the subject index searches for all three topics have a pertinence ratio of 68.00%, whereas the citation index search results have a pertinence ratio of 37.50%. a statistical test affirms the second hypothesis that the difference in the proportions of pertinent references among the unique references retrieved by the two types of index searches is significant. It was found on the basis of the investigation that the subject index search has greater efficacy than the citation search index, since the former scored better with regard to all the measures except redundancy ratio and unit search time. Thus, the citation index search can be a supplementary aid to the subject index search when additional pertinent references are needed. The claim that the citation index gives one the “ease with which one can obtain a negative result with a high degree of certainty” has not been borne out by the data in this research. Nor has the other claim been borne out that a citation index search retrieves pertinent references unexpectedly from fields unfamiliar to the scientist who initiates the literature search. A backward search by checking documents leads to an immense number of references, whereas a forward search leads to a relatively small number of references. The pertinence ration decreases while the noise ration increases with each step of a forward or a backward search. Cycling, therefore, does not appear to be a very promising reference retrieval aid. The ever-increasing noise indicates that it would be useless to do machine searches, either forward or backward, beyond the first step without the elimination of noise at each step. The elimination of noise at each step requires human intervention. The “useful to research” concept should be applied in judging pertinence of references retrieved for topics of extended scope. To know more certainly how often and for what kind(s) of topics the “useful to research” concept needs to be applied in judging the pertinence of references, an analysis of data should be made based on the pertinence judgment both with and without this concept applied in carrying out further research with more topics. How the citing author’s testimony, as checked by the searcher or someone else, compares with the scientist’s pertinence judgment of references is another topic for future research. |
MPACT Scores for Theodore Shih-Shu HuangA = 0 Advisors and Advisees Graphgenerating graph, please reload |
Students under Theodore Shih-Shu Huang
ADVISEES:
- None
COMMITTEESHIPS:
- None