by Christine Neilson
Information Specialist, St. Michael’s Hospital
Toronto, Ontario

Statistics has been on my mind for a variety of reasons lately. As a practitioner/researcher (emphasis on the practitioner part!) I dabble in library research when I can find the time, but I often feel inadequate when it comes to stats. Based on anecdotal evidence, I believe that I’m not alone. I’ve taken introductory stats classes and I know what a p value is, but I feel ill prepared to conduct statistical analysis beyond basic descriptive stats; averages, percentage, and that kind of thing.

The issue of different types of evidence aside, conducting meaningful statistical analysis – correctly – is a matter that has troubled me for a long time. There are a variety of statistical programs available but these tools can’t substitute for actually knowing what you’re doing, and thinking that they will can only lead to trouble. It’s similar to using bibliographic databases without knowing how searching works; thinking that a person should be able to sit down and immediately do an effective, efficient search when they don’t know what the process is, what the commands mean, and when it is appropriate to use which one. But the idea of contracting statistical analysis for a research project to someone else with serious statistical chops somehow seems like cheating. If I’m going to be a real researcher, my internal voice tells me, I should be able to do it myself. I want to be able to do it myself.

Unlike many of the contributors to the Brain-Work blog, I work in a hospital library rather than in academia, and one of my major roles is doing literature searches for health professionals who are conducting research. My colleagues and I provide consults for the do-it-yourselfers, but we encourage our clients to take advantage of our literature search services because we can search better and faster; this isn’t a slight against anyone, it’s simply a fact that we have different areas of expertise. As one of my colleagues likes to say, “Sometimes you just need a librarian”.

So what’s my problem then? We are asking our clients to let go a little bit and trust someone with a specialized skill set, shouldn’t I be able to do the same? If sometimes you just need a librarian, then sometimes you just need a statistician. I’ve been involved in a research team where statistical analysis was delegated to a research assistant with experience doing statistical analysis. A sensible division of labour? Sure. Am I a little relieved that someone who knows what they are doing is in charge of that piece? Honestly, yes. Deep down, do I still want to be able to do it all? You better believe it. But maybe – just maybe – striving for a moderate level of statistical literacy and letting people with more expertise do the heavy lifting might not be such a bad idea after all. I do need to be able to make sense of data analysis when I see it, but whether I like it or not, it is very unlikely that I will have the opportunity to develop real statistical expertise in the foreseeable future.

As I understand it, one of the barriers to librarians conducting research is the intimidation factor. I wonder if more librarians would feel better about the idea of doing research if we embraced the idea that one doesn’t necessarily have to handle every aspect of the endeavour by oneself. Because sometimes you just need a statistician.

This article gives the views of the author and not necessarily the views of St. Michael’s Hospital, the Centre for Evidence Based Library and Information Practice, or the University Library, University of Saskatchewan.