Laws, Sausages, and Statistics

I have just completed the annual task of compiling library statistics for ATS and ATLA a little later than usual, but still before the deadline. I should preface what follows by stating that I like statistics. I like being able to break complex processes down into neat rows of numbers and then lining up those rows of numbers on a spreadsheet. It gives me a secure feeling inside to see all that data marshalled into an eye-friendly format that can be surveyed historically for trends and changes. (Yes, I was one of the popular kids in high school.)

But there is that nagging feeling that an awful lot of ambiguity falls by the wayside when statistics are made. What about the book that comes with full PDF accessibility on a web site for registered purchasers? Did I purchase a book or access to a book? And if I purchased both, how much did each cost? If the publisher helpfully threw in a CD plastered inside the back cover, how do I record that much less circulate it? (Answer: Throw out the CD, and no one will ever know.) Then there are those misleadingly simple numbers about staff and staff salaries. I had one year where a part time professional librarian quit part way through the year. Do I then measure her contribution to staffing totals as 1/4 of a librarian? Or 5/12? The Devil is in the details, which means that there is something diabolical lurking in the white space around those neat rows of numbers.

What is especially troubling to my orderly soul is that we then construct empires out of aggregates of these numbers. "Peer comparison" is one of the benchmarks for how we're doing. It is always much easier to compare yourself to your peer when they are a compact, distant set of numbers, not a flesh and blood institution with crows feet and callused elbows and a host of other minor imperfections that jump out at you. I can say with great confidence that all the libraries in all the ATS schools reported total holdings of 49,280,000 volumes in the 2007-2008 ATS Factbook. (Don't believe me? You can double-check that fact -- I double-dog dare you.) The correlation between that number and the actual number of volumes in all of our libraries ... let's just say I hope that NASA doesn't depend on it to calculate the trajectory for the next Mars mission.

But, like many of my colleagues, I have turned a stony countenance toward those ambiguities, guesses, and WAGs and have now submitted my tidy spreadsheet of numbers to be laid quietly to rest in orderly rows with everyone else's numbers. In addition to being like laws and sausages, statistics are also like cemetaries. It's better not to dig up what lies beneath.