An Examination of Research Collaboration in Psychometrics Utilizing Social Network Analysis Methods
By: Nicole C. DiCrecchio Masters
Published: 5/18/2016
Uploaded: 06/06/2018
Uploaded by: Pocket Masters
Pockets: 2016 (May) Teachers College Columbia University Ed.D. Dissertations, Gottesman Libraries Archive, Historical Dissertations
Tags: Co-author Networks, Longitudinal Study, Psychometrics, Scientific Collaboration, Social Network Analysis

Description/Abstract: Co-authorship networks have been studied in many fields as a way to understand collaboration patterns. However, a comprehensive exploration of the psychometrics field has not been conducted. Also, few studies on co-author networks have included longitudinal analyses as well as data on the characteristics of authors in the network. Including both longitudinal analyses and author characteristics can lead to enhanced understanding of collaboration patterns to determine whether the existing network structure of collaboration in psychometrics hampers or helps the productivity of research. Therefore, this dissertation contributes to the study of scientific collaboration as measured by co-author networks and to the field of psychometrics by including longitudinal analyses, collecting demographic data on the authors as well as a keyword analysis.
Social network analysis methods were used. The data collected came from five prominent psychometric journals over a ten-year period (2005-2014). In order to determine what the network properties were as well as how the psychometrics field has evolved, network statistics were calculated for each year subnetwork and for the overall network. Findings suggest the psychometrics network varied little throughout the time period. Furthermore, the psychometrics network was found to have many small groups of researchers unconnected to the rest of the network, indicating that the field has many subcliques. To explore network structure further, demographic variables such as institution, gender, and country were analyzed to determine how they were associated with collaboration patterns. Results indicated that little collaboration occurred across institutions and across countries. Publication keywords were collected and analyzed to determine if they could identify and explain clustering patterns in the network. Results indicated that many clusters consisted of one productive author and his/her co-authors. This phenomenon tends to create cliques in the network. The keyword analysis also revealed that the topic of equating was an area of increasing interest. The findings have implications for understanding the nature of contributions in the field of psychometrics. In particular, it can be inferred that the field could enhance productivity by making efforts to reduce the occurrence of cliques.