Potatoes to patrons: Using a variation of Fisher’s agricultural split-plot model to explore the information control dimension of LibQUAL+

This research was originally conducted as the author’s dissertation work at the Ohio University. The author explored LibQUAL+ results from two separate institutions with different Carnegie Classifications, and therefore different academic missions, to look for relationships between patron types, Carnegie Classifications, and scores across the minimum, perceived, and desired questions of the information control (IC) component of the LibQUAL+ instrument. By comparing results from a library affiliated with a research institution to one from a campus more focused on teaching and learning, a school going through the shift from one focus to another would be better able to anticipate changes related to patron needs. The paper aims to discuss these issues.

A three-way between-within subjects ANOVA was conducted. The first between-subjects variable was patron type, which included undergraduate, graduate, and faculty. The second between-subjects variable was Carnegie Classification, which included the two classifications of RU_H and Master’s_M. The within-subjects variable had three levels, which in this case functioned as three dependent variables made up of the mean or composite score of the combined eight questions included in the IC portion of LibQUAL+, broken in the three categories of minimum, perceived, and desired. An additional breakdown shows that 499 were undergraduate students, 137 were graduate students, and 197 were faculty.

The results of the study indicated that Carnegie Classification has no significant effect on how undergraduate, graduate, and faculty respond to the three levels of the IC component of the LibQUAL+ survey. As other studies have shown however, there were significant differences with regard to patron-level responses. For a more comprehensive look at all seven research questions and their answers, please see the complete dissertation here: http://rave.ohiolink.edu/etdc/view?acc_num=ohiou1354726349

This study is limited in size and scope because of the limitations of the method of analysis. A broader study using the same analysis would be difficult because of the impracticality of adding, for example, additional Carnegie Classifications into the equation. A significant limitation is that LibQUAL+ results are not typically compared across institutions as the respondents are commenting on separate collections and services. This was minimized by choosing institutions that belong to the same very strong consortial system and have an interlibrary loan system in place which essentially creates one enormous collection for all to share.

Perhaps more significant than the findings themselves is the method of analysis used, as it is one that while complicated statistically, is relatively easy to explain by using the split-plot studies conducted by R.A. Fisher on which the analysis is based as a starting point. The author have found that conceptually it is easier for those without a statistical background to relate to images of potato fields with varying types of potatoes and fertilizer than Carnegie Classifications, patron types, and the multi-level components of LibQUAL+ results.

It would be difficult to speak to the originality of the proposal, but the author would say that a possible outcome would be a discussion of the value of translatable results that speak to broader audiences, particularly those outside library settings. Methods of analysis that can be explained in ways that do not involve the word ANOVA have value and will add to a stronger understanding of research questions and results by decision makers.