| The study investigated research topics of doctoral dissertations that examined issues in distance learning from 2000-2014. Twelve reviews of research on distance learning, spanning from 1997-2015, were identified. It was found that only one of these reviews of research (Davies, Howell, & Petri, 2010) looked at doctoral dissertations. The authors noted that investigating dissertations was complicated and daunting because 1) only a fraction made full text available and 2) there were a large number of dissertations in the area. To counter for these complications the current study utilized bibliometric and social network analysis to investigate dissertation database listings, including abstracts, keywords, classifications, and other bibliographic data.;Bibliographic data for dissertation listings (n=3,954) was exported from the ProQuest Dissertations & Theses A&I (PQDT) database. Software developed for the study formatted the data and imported it into a series of databases. Natural language processing techniques were utilized to pull emergent keywords from dissertation abstracts. Department and University types were analyzed. Dissertation reference sections were investigated utilizing co-citation analysis. Author generated keywords and emergent keywords from abstracts were investigated utilizing keyword co-occurrence network analysis.;Findings indicated that dissertations came from 17 department types including education-oriented department types, such as Educational Leadership, Educational Technology, and Educational Psychology, as well as non-education-oriented departments, such as Business, Psychology, and Nursing. Seven research topics were found to be pervasive in dissertations from 2000-2014: Student, Instructor, Interaction, Administration and Management, Design, Educational Context, and Technological Medium. No change was found over time; rather these seven topics remained the most central nodes in each of the keyword co-occurrence networks. Finally this method of investigation relied heavily on algorithms developed for the study to aid in data formatting and analysis. The merits of this highly automated SNA approach were discussed. Use of abstracts and natural language processing enabled a much higher n size (n=3954) to be investigated than in comparison with the only other study to analyze distance education dissertations Davies et al. (2010) where n=100. This method enabled the heavy lifting to be dedicated to the interpretation of the results, rather than data preparation. |
