by Sue Palmatier, NH State Library
State Data Coordinator

There are lots of ways to lie with statistics:

• forget to count every ILL transaction or every person who attends a program;
• guess how many reference questions have been answered or how many people have entered the library;
• leave blanks on the N.H. Public Library Annual Report.

All of these methods call for estimates - on your part, on my part, or on the part of statisticians at the U.S. Census Bureau.

Estimates that you make are the most likely to be at least close to accurate, since you know your library and your patrons well. Estimates that I make are likely to be less accurate, since I base them on last year's figures - if I have them - and on whatever broad or skimpy knowledge I have of your library. Estimates that Census makes are the least likely to be anywhere near accurate, since they're based solely on statistical probabilities; they don't take into account the 6-inch snowfall that kept people away from Stephen King's appearance at your library, or the flood that wiped out the senior citizen housing next door and sent some of your best borrowers to other towns to live.

So...how can you prevent prevaricating with statistics?

• COUNT. Count every person, every item, every transaction, every square foot, every dollar.
• SAMPLE. You can conscientiously count the following statistics during an "average" week in April and/or October: the library's schedule and hours open (Q. 13 and 14); library visit (Q. 15a); reference questions (Q. 25); volunteer hours (Q. 58); and users of electronic resources (Q. 69). After counting, multiply by 26 or 52 (depending on whether or not you used a week in both April and October) to get the yearly total if that's what the question calls for.
• ESTIMATE. If all else fails, make your best estimate; it's better than having someone outside your library fill in the blanks. Census doesn't even call what they do "estimating"; they call it "imputing" (yes, that's with an "m" and the "u" pronounced as "you"), which according to Webster's means giving responsibility - or blame - to someone else, often falsely or unjustly. Do you really want the Census Bureau employees crediting you with statistics that they created?

But, you ask, do statistics really matter in the long run? We fill out the Public Library Annual Report, send it to the State Library, and everyone forgets it until next year.

Library statistics do matter. They matter when you're thinking about a budget increase for more staff or bigger collections; they matter when you're deciding to automate your catalog and circulation; they matter when you're establishing policies for Internet use or community room occupancy; they matter when you're planning a renovation, an expansion, or a new building.

And your statistics - accurate and timely - matter not just to your library and your community, but to other libraries within New Hampshire and throughout the country. Librarians and trustees in 49 other states, five territories, and the District of Columbia are looking at your statistics on a web site called Public Library Peer Comparison Tool - http://nces.ed.gov/surveys/libraries/publicpeer/. Using one or more variables, any public library in the United States or beyond can compare itself with one or more other libraries that participate in the Federal/State Cooperative System for Public Library Data. Within the last three years, Pennsylvania used data from the Peer Comparison Tool to successfully upgrade libraries statewide through increased state aid and legislative support.

So statistics matter. Be good to yourself and your colleagues by submitting complete and accurate figures on time. Then sit back and have a glass of eggnog, a cup of coffee, a mug of herbal tea, or a nice, relaxing nap. After all, the N.H. Public Library Annual Report won't be back for another year.