圖書資訊領域2006-2010年跨機構知識擴散之研究

Abstract

本研究結合書目計量與網路計量方法，以SSCI之9種圖書資訊領域期刊探討圖書資訊領域近五年（2006-2010）機構之間透過合著、引文顯現的知識擴散現象，並觀察機構於網際網路上的知識與資訊擴散情況。在分析與呈現方式上，本研究運用社會網絡分析與空間分析，分別從社會空間與地理空間討論跨機構的知識與資訊的交流與傳播。 研究結果顯示，在合著方面，跨機構合著為趨勢，2所機構合著為主要型態；合著網絡之核心機構為北卡羅來納大學（University of North Carolina）資訊與圖書館學學校和馬里蘭大學（University of Maryland）資訊研究學院。在引文方面，跨機構引用數逐年遞增，且有頻繁引用特定機構的情況；引文網絡之核心機構為印地安那大學（Indiana University）圖書館與資訊科學學校和北卡羅來納大學資訊與圖書館學學校。在網路連結方面，機構外連數差異懸殊，核心機構為美國國立衛生研究院（National Institutes of Health, NIH）美國國家醫學圖書館、美國國會圖書館、賓州州立大學（Pennsylvania State University）資訊科學與科技學院。無論合著、引文、網路連結，其關係網絡各別皆由頻繁互動機構串連組成，而以強成分劃分之引文與網連網絡更可用以解釋知識擴散。此外，合著、引文、網連機構知識擴散於地理空間上各別皆呈群聚分佈，機構合著與引文高密度區為英國、美國，網路連結為美國。 在合著、引文、網路連結的整體比較方面，本研究發現合著與引文擴散網絡呈正相關，網連與其他兩者則無相關；而網連為三者網絡中密度較高者，但合著與引文密度隨時間增加；此外，三者網絡主要機構國別為美國，機構類型於合著與引文網絡多為學院、網連則為圖書館，北卡羅來納大學和賓州州立大學的圖資系所整體而言對圖資領域知識擴散最具貢獻。另研究發現，三者知識擴散熱區為美國，且皆以擴散至鄰近機構為主，網連擴散距離最遠，合著與引文隨時間漸突破空間限制。

This study combined Bibliometrics with Webometrics to explore knowledge diffusion through coauthorship and citation among institutions of Library and Information Science (LIS) in recent five years (2006-2010) by examining LIS institutions from 9 LIS journals in SSCI database, and to observe knowledge and information diffusion among institutions on websites (web linkage). Social Network Analysis and Spatial Analysis were applied to measure the knowledge interflow and dissemination among institutions in this study. The major findings were summarized as follows. The tendency of inter-institution coauthorship was increasing, and the main pattern of coauthorship was two-institution collaboration. The core institutions in the coauthorship network were the School of Information and Library Science, University of North Carolina (UNC), and the College of Information Studies, University of Maryland (UMD). In the aspect of citation, the number of inter-institution citation has grown over time with a frequent intension citing specific institutions. The core institutions in citation network were the School of Information and Library Science, University of North Carolina, and the School of Library and Information Science, Indiana University. On the perspective of web linkage, the outdegrees of institutions differed significantly. The core institutions in web linkage network were the United States National Library of Medicine, National Institutes of Health (NIH), the Library of Congress (US), and the College of Information Sciences and Technology, Pennsylvania State University. For the three methods (coauthorship, citation, and web linkage), the networks were all formed by institution-pairs that cooperated more, cited more, or linked more; meanwhile, the subgroups divided by strong component analysis could provide higher explanatory power to the diffusion of the networks of coauthorship and citation. In addition, it has been identified that the spatial distribution patterns of the three methods were all clustered in certain geographical space; moreover, the high-density areas of coauthorship and citation were both in England and the United State, while the web linkage was in the United State. Compared among the three methods, the correlation between diffusion networks of coauthorship and citation appeared positive, but the network of web linkage had no correlation with the other two. On the other hand, the web linkage network had relatively higher connection density than the other two networks, but the densities of the other two networks had grown over time. Furthermore, the core institutions of the three methods’ networks were all in the United State. As to the type of core institutions, the main core institutions of both coauthorship network and citation network were schools or colleges, but those of web linkage network were libraries. The most contributed institutions in LIS knowledge diffusion were the LIS schools in the University of North Carolina and the Pennsylvania State University. Findings also revealed that the three methods’ hot spots of knowledge diffusion were all in the United State, and spreaded to geographically-adjacent institutions. Among them, web linkage had the longest average distance, but the other two could break the limitation of geographical space gradually with time.