生醫科學期刊引用數據之分析比較研究

An Analysis and Comparison of Citation among Journals of Biology, Medical Science, Agriculture and Food Science

蔡明月(Ming-Yueh Tsay) ； 陳憶玲(Yi-Ling Chen)

《國家圖書館館刊》 95年1期 (2006/06) Pp. 1-33

摘要

本研究旨在探討生醫科學類期刊的引用數據特性，以生物學、醫學、農業與食品科學三門學科為研究對象，透過各學科最新參考工具指南所列重要期刊，與美國ISI公司出版之網路版「期刊引用報告(JCR)」2002年「Science Edition」資料庫比對後，確定核心期刊樣本。根據JCR提供各期刊之刊載文章篇數。期刊被引用次數、影響係數、即時引用指數、引用半衰期及被引用半衰期等六項引用數據深入分析，探討各學科期刊基本資料及其引用數據特性；進而利用皮爾森積差相關係數，檢測引用數據間的相關性；並採用費雪爾(Fisher) Z轉換，檢測引用數據相關程度之差異性；再進行獨立樣本T檢定，考驗引用數據之同質性及其平均數差異性。本研究期刊樣本計有：生物學327種、醫學129種、農業與食品科學130種。研究結果歸納如下：
生醫科學期刊大多是月刊、雙月刊或季刊形式。生物學年出刊頻率最多樣，達27種。年刊載1至99篇之期刊所佔比例最高，其次為100至299篇；刊載300至1,000篇者，以醫學為最多。被引用1,000至9,999次的期刊，佔所有被引用次數的比例最高，被引用次數最高者為醫學，主要集中在10,000至49,999次。各學科期刊之影響係數在0.001至1.999所佔的比例最多，生物學、醫學二學科期刊影響係數普遍較高。期刊即時引用指數多集中在0.001至0.199，其中生物大於醫學；醫學又大於農業與食品科學。生物與醫學期刊引用半衰期大多介於5至7年，農業與食品科學則集中在10年以上，顯見其引用資料較不重視時效性。大於10年的被引用半衰期，以農業與食品科學最多，其次為醫學，再次為生物，顯示生物學文獻老化較快。三門學科引用數據正相關度最高者為影響係數與即時引用指數、刊載文章篇數與被引用次數，其次是期刊刊載文章篇數與被引用次數、被引用次數與即時引用指數，被引用次數與影響係數。影響係數與引用半衰期間，生物學、醫學二學科為中度負相關；農業與食品科學為低度負相關。相關度再次者則為被引用半衰期與引用半衰期，生物與醫學二學科均為中度正相關。醫學與農業與食品科學之期刊刊載文章篇數與影響係數，為低度正相關。就三學科間引用數據相關程度的差異性而官，期刊生產量與被引用次數、被引用次數與影響係數、影響係數與即時引用指數的差異性最大；影響係數與引用半衰期、即時引用指數與引用半衰期、被引用半衰期與引用半衰期、被引用次數與即時引用指數則分別有二個學科的差異性達顯著水準。三學科相關程度均無差異的引用數據有二種，分別是刊載文章篇數與即時引用指數、被引用次數與被引用半衰期。

關鍵字

生醫科學期刊 ； 期刊引用報告 ； 期刊引用分析 ； 期刊生產量 ； 被引用次數 ； 影響係數 ； 即時引用指數 ； 引用半衰期 ； 被引用半衰期

並列摘要

The purpose of this study is to analyze and compare journal citation data from Journal Citation Reports (JCR), and focus on three subjects, i.e., biology, medical science, agriculture and food science, on life science. The journal samples were drawn from the core journal list of Guide to Information Sources and ”JCR science edition” on the Web 2002 in biology (327 records), medical science (129 records), agriculture and food science (130 records). The source items and five types of citation data, including citation counts, impact factor, immediacy index, citing half-life and cited half-life, are examined, and the correlation between each of the fifteen pairs of citation data is determined based on the Pearson correlation tests. The Fisher's Z-transform is employed to test the significant difference between the Pearson correlation coefficient for each pair of citation data of these three subject areas. Moreover, the T-test is used to determine the probability that the difference in the means of each citation data being observed is significantly different from zero. The results of this study reveal that:
Most of the life science journals are issued monthly, bimonthly, or quarterly; the publication frequency of biological journals is various. The journals that published 1 to 99 papers are the largest group; whereas journals that published 100 to 299 papers belong to the second group. Most of the medical journals publish 300 to 1,000 articles yearly. Most of the journals are cited 1,000 to 4,999 times; medical journals are cited more than other subjects, the cited times aggregated between 10,000 and 49,999. 0.001 to 0.999 and 1.000 to 1.999 are two largest impact factor categories. Generally, biological and medical journals have higher impact factor. The immediacy index is mostly concentrated on 0.001 to 0.199. In average, immediacy index of biological journals is greater than medical journals and medical journal is greater than agriculture and food science. Citing half-life is focused on 5 to 7 years, most of the agriculture and food science journals have citing half-life more than 10 years. Most of the biological (4 to 7.9 years) and medical (7 to 7.9 years) journals obviously have shorter cited half-life than that of agriculture and food science journals (＞10 years).
Journals with higher impact factor may also receive higher immediacy index. The more papers published yearly the more citations the journal received. The correlations between source items and citation counts, between citation counts and impact factor and between citation counts and immediacy index are positive. Biological and medical journals have middle negative correlation between impact factor and citing half-life. In agriculture and food science journals, there is lower negative correlation. Journals of biological and medical fields reveal middle positive correlation between citing and cited half-life. The lowest correlation is between source items and impact factor. Medical and agriculture and food science journals have lower positive correlation. For journals of three subjects, there is significant difference of the Pearson correlation coefficient as the following pair of citation data: source items and citation counts, citation counts and impact factor, and impact factor and immediacy index; there are no significant difference of the Pearson correlation coefficient for source items and immediacy index and for citation counts and cited half-life. Impact factor has the significant mean difference in three disciplines, followed by immediacy index and citing half-life. The mean difference between biological and agriculture and food science journals is significant. The distinction of mean difference between journal citation's data is influenced by cited half-life and citing half-life, especially between biological and medical journals. There is no difference between the mean difference in source item and impact factor.

並列關鍵字

Journal Citation Data ； Life Science Journals ； Source Items ； Citation Counts ； Impact Factor ； Immediacy Index ； Citing Half-life ； Cited Half-life ； Pearson Correlation Coefficient ； Fisher's Z-transform ； T-test ； JCR