|Abstract: ||本研究以國內某科研機構圖書館於2011 年與2012 年所購七種西文電子期刊庫為例，採用使用統計分析、主題分析及影響係數比較分析，探討電子期刊被收錄變動與複本情形、全文被下載情形、主題分布及下載次數排名與影響係數排名之相關性。主要目的作為電子期刊選購決策、館藏發展及改善服務品質的參考，以及瞭解期刊影響係數作為期刊選購評估要素之可行性。|
研究結果顯示：專業學會及出版社類型資料庫收錄期刊之變動性較小，變動率介於0% 至12.50% 之間，匯集商資料庫收錄期刊變動性大，EBSCO M&G 高達67.86%。僅0.19% 之期刊被不同資料庫重複收錄。在主題分布方面，期刊主題分屬於227 類學科領域，電子與化學研發領域相對應之期刊主題類別數與期刊數最多，兩者合計佔全部類別數23.76%，佔全部期刊數30.13%；其他非研發領域直接相關之期刊主題類別數與期刊數所佔比率偏高，佔全部類別數41.35%，佔全部期刊數31.05%，大部份為套裝方式購入。ScienceDirect 與IEL 兩種資料庫包含與科研機構研發領域直接相關之期刊數及全文下載次數較多，兩者合計佔全部期刊數57.19%，佔全部下載次數89.49%，SPIE 資料庫則是平均每種期刊全文被下載數最高151 次。在使用與影響係數相關性方面，個案所購電子期刊之全文下載次數排名與其影響係數排名呈現低度負相關。
This study used seven electronic journal databases in English purchased by an anonymous domestic library in 2011 and 2012 as examples. Statistical analysis, thematic analysis, and comparative analysis of impact factor were adopted to explore the situation of changes in electronic journal excerpts and duplicates, situation of full-text download, theme distribution, as well as the relevance of download frequency ranking, and impact factor ranking. The main purpose of this study is to serve as a reference for decision-making pertaining to electronic journal purchases, library collection development, and service quality improvement in order to gain an insight into the feasibility of using journal impact factor as factors for journal purchase evaluation.
The results of the study reveal that:
1. Electronic journal excerpts from professional societies and publisher databases had a smaller rate of change, between 0% and 12.50%. The journal excerpts of aggregators’ databases had a larger rate of change, EBSCO M&G accounting for 67.86%.
2. While only 0.19% of the journals were excerpts that repeatedly appeared in different databases.
3. In terms of theme distribution, the journal themes fell under 227 types of disciplines, with the journal themes and quantity of journals corresponding to electronic and chemical R&D fields accounting for the majority, 23.76% of the total quantity of categories and 30.13% of the total quantity of journals.
4. The quantity of journal theme categories and the quantity of journals directly related to non-R&D fields were relatively higher, accounting for 41.35% of the total quantity of categories and 31.05% of the total quantity of journals. Most of the journals were purchased in packages.
5. The ScienceDirect and IEL were two databases that covered more quantity of journals directly related to the R&D field of scientific and research institutions, accounting for 57.19% of the total quantity of journals and 89.49% of the total download frequency. On the other hand, the SPIE database had the highest average number of full-text download for each journal type, accounting for 151 times.
6. As for usage and impact factor relevance, the full-text download ranking and impact factor ranking of the electronic journals purchased by the cases showed a low degree of negative correlation.
Based on the research results, it is suggested that the cases apply the results to develop purchase strategies, adjust the library collection excerpt scope, and actively engage in promotion and marketing targeting electronic journals with low usage rates, in order to avoid using journal usage rates or impact factors as the only standards for eliminating journals. Libraries should also take the various factors into consideration, take a supportive role in scientific research institutions’ research and development, and establish journal collections.