Van Raan (1997)討論科學計量學的最佳狀態，強調這個領域需要平衡應用與基礎研究以及增強科學計量學和廣泛的學科之間的關係。過去已經有許多研究根據Scientometrics期刊論文的參考文獻，對科學計量學領域的特性進行分析，例如：Schubert and Maczelka (1993)利用1980-81與1990-91兩個期間的論文參考文獻，分析1980年代Scientometrics期刊的改變，所使用的指標包括參考文獻的年齡分布(age distribution)、Price指標(Price index)、引用出版品的分布、引用作者的分布以及最常引用的出版品等。Wouters and Leydesdorff (1994)分析前25卷的論文參考文獻，計算論文的平均參考文獻數量以及被引用文獻的相對年齡(relative age)。Schoepflin and Glänzel (2001)以1980、1989和1997年的論文參考文獻資料，計算Price指標、參考文獻為連續出版品(serial)的比率、參考文獻的平均年齡(mean reference age)和平均引用的參考文獻比率(mean reference rate)。
本研究比較Scientometrics期刊1990和2000年論文的參考文獻，共169篇論文(1990年70篇，2000年89篇。附註：此處有誤，合計只有159篇。)，2814筆參考文獻，並將參考文獻的來源分為不同領域。結果發現1990年每篇論文平均有15.1(1054/70)筆參考文獻，2000年每篇論文平均有19.8(1760/89)筆參考文獻。Scientometrics期刊論文引用近五年的文獻比例，也就是Price指標，從1990年的37.6%，減少為2000年的31.6%。1990年有47.3%的參考文獻來自科學計量學與書目計量學(scientometrics and bibliometrics)、圖書資訊學(library and information science)以及社會學、歷史學與哲學(the sociology, history and philosophy of science)，2000年增加為56.9%。兩個年度的作者自我引用情形沒有明顯差異，1990年為13.4%，2000年則為13.9%。然而該期刊的自我引用(journal self-citation)與引用期刊的百分比(the percentage of references to journals)等情形皆有增加。期刊的自我引用情形從1990年的12.9%增加為2000年的20.1%，可能的理由是Scientometrics愈來愈成為這個領域的核心期刊。
The aim of this study was to examine the extent to which the field of bibliometrics and scientometrics makes use of sources outside the field.
The results show that in 2000, 56.9% (and 47.3% in 1990) of the references originated from three fields: scientometrics and bibliometrics; library and information science; and the sociology, history and philosophy of science.
When comparing the two periods, there is also a considerable increase in journal self-citation (i.e., references to the journal Scientometrics) and in the percentage of references to journals.
Van Raan (1997) discussed the state-of-the-art of scientometrics, and emphasized the need to balance between applied and basic research in the field, and the importance of strengthening the relations of scientometrics with a broad spectrum of disciplines.
Peritz (1981) examined the references of research papers published in 39 core journals during five calendar years, and calculated the percentage of references outside the field. In another study, Peritz (1988) examined the literature for bibliometrics for the period 1960–1985, and classified the body of literature according to the field of the journal in which the article was published.
Al-Sabbagh (1987) studied the interdisciplinarity of information science through a reference analysis of JASIS. The findings, based on a ten percent random sample of references appearing in JASIS articles between 1970 and 1985 show that the largest percentage of references come from information science, followed by computer science, library science and science-general.
Thompson (1989), based on the references of articles in twenty library and information science journals in five selected years, studied the age of the references, the extent of self citation, and found that the list of most cited journals was almost exclusively from library and information science.
Cronin and Pearson (1990), in a study based on citations found that information science exports techniques of information retrieval and bibliometrics.
Meyer and Spencer (1996) analyzed citations to twenty-four library and information science journals over a twenty-year period. Their findings show that 86.6% of the citations come from library and information science, but other disciplines including computer science, medicine, psychology, the social sciences and general sciences also cite library and information science journals to some extent.
Rousseau (1997) studied the references appearing in the papers of the first two ISSI Conferences and citations of the Proceedings. He tabulated the most frequently cited publications – the three most frequently cited publication were JASIS, Scientometrics and J. Doc. The list of most frequently citing journals and of the most cited papers from the Proceedings was also presented.
The journal Scientometrics has been the “theme” of several previous bibliometric studies.
Schubert and Maczelka (1993) studied the changes that occurred to the journal during the 1980’s based on its reference patterns. Two periods, 1980–81 and 1990–91 were chosen. They calculated the age distribution of the references, the Price Index, the distribution of the cited publication, the distribution of cited authors, and tabulated the most frequently cited publications.
Wouters and Leydesdorff (1994) analyzed the references of all articles and notes in the first 25 volumes of Scientometrics, and calculated, among other indicators, the number of references per article and the relative age of the cited literature (the Price Index).
Persson (2000) maps the citation and reference patterns of Scientometrics based on volumes 1 to 44 of the journal.
Very recently Schoepflin and Glänzel (2001) calculated several bibliometric measures (the Price Index, percentage of references to serials, mean reference age and mean reference rate) for articles, letters and notes published in Scientometrics in 1980, 1989 and 1997.
Most of the previous studies of Scientometrics were either concerned with quantitative aspects (e.g.; the Price Index) or with citation and co-citation patterns.
This study analyzed all the references of all the papers published in Scientometrics in 1990 and in 2000. The population of the study consisted of 169 papers and 2814 references. ...The list contained 70 items for 1990 and 89 items for 2000, altogether 169 papers. ... In the set for 1990, 60 (86% of the total for 1990) items in the list were labeled as articles, while for 2000, 83 (93% of the total for 2000) items were labeled as articles.
The references were categorized according to six facets:
• author self-citation;
• journal self-citation;
• discipline of publication source;
• field self-citation;
• type of publication;
• year of publication.
Thus, first we had to define the major themes covered by the field:
• indicators (science and technology), forecasting and planning;
• research trends, research evaluation and funding;
• science policy;
• bibliometric laws and models;
• citation analysis including all aspects (e.g. obsolescence, ranking, mappings, coupling, etc.);
• patent analysis;
• reference analysis;
• coword analysis in context of performance;
• productivity (e.g. authors, journals, institutions);
• impact;
• peer review process;
• sociology of science;
• social contexts of research;
• characteristics and development of a scientific area;
• scholarly communication;
• scientific networks;
• technology flow;
• innovation;
• other themes relevant to the field.
Altogether, 2814 references were identified, 1054 in 1990 and 1760 in 2000.
The mean number of references in 1990 was 15.1, while in 2000 the mean increased to 19.8.
It is interesting to note, that the percentage of references in the last five years (1986 to 1990, for 1990; and 1996 to 2000 for 2000) decreased from 37.6% to 31.6%. Does this mean that scientometrics is getting “softer” (the “Price Index”, (Price, 1970))?
Both Schubert and Maczelka (1993) and Shoepflin and Glänzel (2001) found that the Price Index of scientometrics increased over time. They, as in the current work, based their data on single years.
On the other hand, Wouters and Leydesdorff (1994) studied the first twenty five volumes of Scientometrics, observed some fluctuations in the Price Index over the years, but showed that the regression line is not significant, and concluded that the index displays neither rise or fall between 1978 and 1992.
We also believe, that in order to draw conclusions about the “hardness” or “softness” of the field, its journal or journals should be studied over a continuous time period, and not isolated years.
A reference for a given item was labeled as author self-citation, if one of the authors of the reference matched one of the authors of the given item. ... Author self-citation was 13.4% for 1990 (141 references) and 13.9% (244 references) for 2000. ... In terms of author self-citation, no significant differences were observed between the two periods.
Journal self-citation, (i.e., references to the journal Scientometrics), on the other hand, increased considerably, from 12.9% in 1990 (136 journal self-citations) to 20.1% (354 journal self-citations) in 2000.
A possible explanation for this increase is that the journal Scientometrics is more and more becoming the central journal of the field.
The top four journals appear exactly in the same order for both years, and they represent the main aspects of the field: the field itself, its relation to information and library science, to planning and management and to the sociology of science. These four sources cover 18.5% of the references in 1990, and 28.1% of the references in 2000.
In the list of most frequently cited publications we see mostly journals, but also books, handbooks, yearbooks, collections, proceedings and reports. ... Table 4 shows that the percentage of the references to journal articles increased considerably, while the percentage of the references to books, yearbooks and reports decreased. In 1990 there were no references to electronic sources, this category only appeared in 2000. It will be interesting to see whether the electronic sources are going to be referenced more extensively in the future.
Along with the increasing citation rate of the journal Scientometrics, we observe a general increase in sources belonging to the field of scientometrics and bibliometrics. This could either be the sign that the field is becoming more mature or self-sufficient or it may indicate that scientometricians base their research less and less on methods and studies conducted in other fields.
Most of the references 2000 (56.9%), and nearly half of the references in 1990 (47.3%) are from the three fields, closely related to the subject-matter: scientometrics and bibliometrics itself; library and information science; and sociology and history of science.
A substantial amount of references are to sources belonging to the social sciences (21.3% in 1990, and 13.0% in 2000). We see that the percentage of references to sources from the social sciences decreased considerably.
On the other hand, the combined share of sciences-general; mathematics, computer science, statistics and engineering; science and medical sciences remained nearly the same (23.4% in 1990 versus 23.5% in 2000).
About half of the references originate from sources, which are not related to scientometrics. It is quite possible that some of these references are to works, which belong to the field.
On the other hand, some of the references from the fields closely related to scientometrics are not classified as field self-citation.
In 1990, 593 out of the 1054 references (56.3%) were classified as field self citation, while in 2000, 1092 out of the 1760 references (62.0%) were field self-citations. This is a rather considerable increase, it may indicate that the field is becoming more and more self sufficient, and needs to rely less on theories and methods emanating from other scientific fields.
The results show that the field relies heavily on itself, on library and information science and on sociology, history and philosophy of science.
There is an increase in journal self-citation, the list of core journals remaining stable for both periods. Author self-citation is around 20% for the years under study.
Most of the previous studies of Scientometrics were either concerned with quantitative aspects (e.g.; the Price Index) or with citation and co-citation patterns.
This study analyzed all the references of all the papers published in Scientometrics in 1990 and in 2000. The population of the study consisted of 169 papers and 2814 references. ...The list contained 70 items for 1990 and 89 items for 2000, altogether 169 papers. ... In the set for 1990, 60 (86% of the total for 1990) items in the list were labeled as articles, while for 2000, 83 (93% of the total for 2000) items were labeled as articles.
The references were categorized according to six facets:
• author self-citation;
• journal self-citation;
• discipline of publication source;
• field self-citation;
• type of publication;
• year of publication.
Thus, first we had to define the major themes covered by the field:
• indicators (science and technology), forecasting and planning;
• research trends, research evaluation and funding;
• science policy;
• bibliometric laws and models;
• citation analysis including all aspects (e.g. obsolescence, ranking, mappings, coupling, etc.);
• patent analysis;
• reference analysis;
• coword analysis in context of performance;
• productivity (e.g. authors, journals, institutions);
• impact;
• peer review process;
• sociology of science;
• social contexts of research;
• characteristics and development of a scientific area;
• scholarly communication;
• scientific networks;
• technology flow;
• innovation;
• other themes relevant to the field.
Altogether, 2814 references were identified, 1054 in 1990 and 1760 in 2000.
The mean number of references in 1990 was 15.1, while in 2000 the mean increased to 19.8.
It is interesting to note, that the percentage of references in the last five years (1986 to 1990, for 1990; and 1996 to 2000 for 2000) decreased from 37.6% to 31.6%. Does this mean that scientometrics is getting “softer” (the “Price Index”, (Price, 1970))?
Both Schubert and Maczelka (1993) and Shoepflin and Glänzel (2001) found that the Price Index of scientometrics increased over time. They, as in the current work, based their data on single years.
On the other hand, Wouters and Leydesdorff (1994) studied the first twenty five volumes of Scientometrics, observed some fluctuations in the Price Index over the years, but showed that the regression line is not significant, and concluded that the index displays neither rise or fall between 1978 and 1992.
We also believe, that in order to draw conclusions about the “hardness” or “softness” of the field, its journal or journals should be studied over a continuous time period, and not isolated years.
A reference for a given item was labeled as author self-citation, if one of the authors of the reference matched one of the authors of the given item. ... Author self-citation was 13.4% for 1990 (141 references) and 13.9% (244 references) for 2000. ... In terms of author self-citation, no significant differences were observed between the two periods.
Journal self-citation, (i.e., references to the journal Scientometrics), on the other hand, increased considerably, from 12.9% in 1990 (136 journal self-citations) to 20.1% (354 journal self-citations) in 2000.
A possible explanation for this increase is that the journal Scientometrics is more and more becoming the central journal of the field.
The top four journals appear exactly in the same order for both years, and they represent the main aspects of the field: the field itself, its relation to information and library science, to planning and management and to the sociology of science. These four sources cover 18.5% of the references in 1990, and 28.1% of the references in 2000.
In the list of most frequently cited publications we see mostly journals, but also books, handbooks, yearbooks, collections, proceedings and reports. ... Table 4 shows that the percentage of the references to journal articles increased considerably, while the percentage of the references to books, yearbooks and reports decreased. In 1990 there were no references to electronic sources, this category only appeared in 2000. It will be interesting to see whether the electronic sources are going to be referenced more extensively in the future.
Along with the increasing citation rate of the journal Scientometrics, we observe a general increase in sources belonging to the field of scientometrics and bibliometrics. This could either be the sign that the field is becoming more mature or self-sufficient or it may indicate that scientometricians base their research less and less on methods and studies conducted in other fields.
Most of the references 2000 (56.9%), and nearly half of the references in 1990 (47.3%) are from the three fields, closely related to the subject-matter: scientometrics and bibliometrics itself; library and information science; and sociology and history of science.
A substantial amount of references are to sources belonging to the social sciences (21.3% in 1990, and 13.0% in 2000). We see that the percentage of references to sources from the social sciences decreased considerably.
On the other hand, the combined share of sciences-general; mathematics, computer science, statistics and engineering; science and medical sciences remained nearly the same (23.4% in 1990 versus 23.5% in 2000).
About half of the references originate from sources, which are not related to scientometrics. It is quite possible that some of these references are to works, which belong to the field.
On the other hand, some of the references from the fields closely related to scientometrics are not classified as field self-citation.
In 1990, 593 out of the 1054 references (56.3%) were classified as field self citation, while in 2000, 1092 out of the 1760 references (62.0%) were field self-citations. This is a rather considerable increase, it may indicate that the field is becoming more and more self sufficient, and needs to rely less on theories and methods emanating from other scientific fields.
The results show that the field relies heavily on itself, on library and information science and on sociology, history and philosophy of science.
There is an increase in journal self-citation, the list of core journals remaining stable for both periods. Author self-citation is around 20% for the years under study.
沒有留言:
張貼留言