The Effects of Reference, Instruction, Database Searches, and Ongoing Expenditures on Full-text Article Requests: An Exploratory Analysis

• Published in: The Journal of academic librarianship 2014-05, Vol.40 (3-4), p.264-274
• Main Author: Fagan, Jody Condit
• Format: Article
• Language: English
• Subjects: Academic libraries; ACRL Trends & Statistics Survey; Communication. Information transfer; Data base management systems; Database searches; Enrollment; Exact sciences and technology; Expenditures; Information and communication sciences; Information science. Documentation; Information service management; Interlibrary loans; Library and documentation centre management; Library and information science. General aspects; Library assessment; Library Instruction; Library metrics; Multiple regression; Regression (Statistics); Sciences and techniques of general use; Searches; Trends; User service management
• ISSN: 0099-1333; 1879-1999
• DOI: 10.1016/j.acalib.2014.04.002

ACRL Library Trends and Statistics survey data from two years were analyzed to explore relationships among variables thought to predict full-text article requests. Five different regression models predicted full-text article requests from variables including reference transactions, library instruction, Web site visits, database searches, and ongoing expenditures. Full-time student enrollment and Carnegie classification were included as covariates in all five models. All regressions supported using enrollment as a covariate; four of the regressions supported using Carnegie classification as an additional covariate. Results suggested that reference transactions, library instruction, and ongoing expenditures increase the number of database searches and full-text requests beyond these covariates. Ongoing expenditures made the largest contribution in all models, but did not always contribute the most unique variance. Furthermore, a mediation analysis showed a significant indirect effect of library instruction on full-text requests through database searches. This study suggests most ACRL data may need log transformation before using regression analyses, and offers one approach to interpreting results. Future research could investigate the meaning of ongoing expenditures after controlling for institutional size and Carnegie classification. Additional variables such as interlibrary loan usage or whether the library has a discovery tool might be interesting to include in similar models.