The purpose of this paper is to discuss the importance of selecting “well-matched” independent and dependent variables in quantitative research to maximize the possibility of detecting impact of library services on indicators of student success. The paper introduces the concept of sensitivity, which is the extent to which a measure will detect change in the thing being measured.
To make the case, the authors use the impact of amount of library instruction received on Grade Point Average (GPA) as an example, explaining a correlational research study at their institution. However, the emphasis of the paper is on the conceptual importance of sensitivity in variable selection in quantitative studies of all types.
After finding no statistically significant relationship between the amount of library instruction received and GPA, the authors determined that GPA was not a sensitive enough variable to detect the impact of a few class sessions taught by a librarian throughout students entire undergraduate career. Based on the findings and the literature, the authors conclude that the practice of selecting “insensitive” dependent variables that are unlikely to detect impact of the independent variable is a common practice in the library assessment literature.
In an era where bigger is better when it comes to demonstrating impact of library services, this paper argues that libraries sometimes diminish their ability to illustrate their contributions to student success by choosing large scale indicators of student success as independent variables which fail to detect the impact of library services.
