What Counts? Using Data to Influence Practice

Presented by Tonnie Flannery and Jessica Johanningmeier of Cornell College

While many are fond of using rigorous statistical methods, sometimes for making rather unimpressive distinctions, the authors of this paper

show that fairly simple descriptive measures can yield very useful results and inform the process of asking the right questions.

In the first case, an unrestricted interlibrary loan policy generated a very large and increasingly expensive mass of requests. The tracking system at first produced only a distinction between requests for photocopies and requests for original materials (loans) which did not prove particularly helpful in explaining what was happening.  Recoding the raw data into spreadsheet form allowed creation of a much clearer picture of the patterns of use, types of materials called for, timing of requests, and variations of use within different sectors of the user population.  More detailed inspection proved that much of the demand was for entertainment-related CDs and videos.  Other measures showed patterns of demand that were generated by class use and those that appeared unrelated.  The analysis allowed for a reasoned restriction of the ILL policy to meet budgetary necessities without unduly limiting requests needed for classroom and research purposes, and also revealed problems in searching that could be remedied during user-education sessions.

A second project concerned use of the Quantitative Reasoning Studio. Again, studying patterns of use revealed how to most efficiently target resources. Addition of a few questions to user surveys allowed tracking of use by freshmen vs. juniors, for example, and also comparing usage trends between, say, freshmen of one admission year and those of another.

What seems to be most significant is that both projects were of a practical nature related to improving service at a specific institution and for the user populations involved. In both cases the methods used allowed detection of broad trends that were not immediately obvious from the raw data.  Perhaps as useful were the relatively easy means of developing new questions to ask of the data and the ingenuity employed in finding practical applications of the results beyond what was initially sought.

Powerpoint available.

Submitted by Jeff Dodd, Newsletter Committee