IRIS Photos

Reading the Tea Leaves: Examining CUL Circulation Statistics

Anne R. Kenney

Like many of you, I recently received the CUL Annual Statistics 2003/2004 (PDF). As I flipped through the pages, the statistical graph on library usage struck me. Circulation had declined from the previous year. This decline came on the heels of dramatic increases in circulation in the previous two years, in large measure attributed to the recon effort. Did the chart reflect the Cornell version of the general downward trend in circulation that research libraries across the country are reporting? Digging a little deeper, I reviewed the circulation statistics for the individual libraries (Table 7a) and, when I dropped the laptop circulation figures from the totals, the overall decline was 5%, but nine libraries actually experienced an increase in circulation, eight saw declines, and one (Geneva) remained the same. What really struck me, however, was the decline in Uris Library’s stats—a 47% drop in circulation from one year to the next. In other words, 102,629 fewer items circulated in 2003/04 than in 2002/03.

What was going on here? Were books really on their way out, as my son keeps proclaiming? Will some senior administrator take a look at these charts and conclude that university funds can be redirected elsewhere? Would that person be right? What, in fact, do the statistics really tell us? With a lot of help from my friends (thanks especially to Carmen Blankinship, Barbara Eden, Duane Edwards, Linda Miller, Lydia Pettis, and Pat Schafer), I dug deeper.

My first thought was that the statistics were just wrong, but no, a recheck indicated their accuracy—automatically generated statistics are usually fairly reliable. Thinking back, I recalled that the media collections had been relocated to Olin from Uris, which could account for some of the difference. Lydia and Duane kindly ran reports to isolate the media circulation figures. Last year Uris’s media figures dropped from 40,822 to 11,201 (mostly reserve), a decline of 29,681, while Olin’s media figures rose from 2,311 to 54,677—an increase of 52,366. The shift of media from Uris to Olin could account for 29% of the total drop in circulation at Uris. More significant is that media use is on the rise (increasing 53% in one year for Olin/Uris combined).

It was then that Pat suggested there might be something in the way laptops were checked out in 2002/03. Although I had backed out the laptop figures, laptop peripherals were also barcoded and checked out separately. In fact, up to four circulation transactions could have been recorded for each laptop checkout. In 2003/04, the laptops and peripherals were counted as one circulation, although if a carrying case was also needed, it was counted as a separate circulation figure. A check of three years’ worth of statistics indicated that 2002/03 was the anomalous year. Lydia ran another report comparing monthly charges/renewals in Uris from 2000 to the present, which indicated that the surge in Uris circulation activity began in September 2002 and continued through June 2003. Uris circulation figures in 2002/03 rose by 52% over 2001/2002, with 74,924 more circulations. Suppose all those transactions represented computer peripheral charges. If you added them to the drop in media circulation, the figure would be 104,605 and would more than account for the circulation decline from 2002/03 to 2003/04. It’s theoretically possible then, that Uris’s book circulation figure actually increased during this timeframe! Are there similar explanations for the drops in circulation at Hotel, Music, Mann, and Engineering? Obviously, one has to be careful in drawing firm conclusions from the Annual Statistics.

Statistics rarely lie, but they can greatly deceive, and we should look more carefully at what yardsticks we use to measure library services. In fact, our current statistics inadequately capture our biggest growth industry—online access to resources. Table 7c in the Annual Statistics represents electronic reserves in terms of the number of courses, articles, and article views. By all three counts e-reserve use is on the rise. In Uris, for example, the number of article views increased by 141% in two years from 70,865 in 2001/02 to 171,134 in 2003/04. That’s a significant use figure that is not added to the circulation figures or represented in the statistical graph on library usage (p. 25 in the report). Table 8 reports Library Gateway hits, but anyone who has tried to rely on hits knows how little they actually correlate to measurable use. And what about use of the Making of America Collection (which tallies close to 2,000,000 page views/month) and our other online digital content? It’s no wonder that Sarah Thomas has asked Research and Assessment Services to work with library staff to get a better handle on use statistics associated with e-resources. Their findings should inform the types of use data we collect and, equally important, how we interpret and present this information to various constituencies.

In sum, the moral of this story is to take a grain of salt with each reading of the tea leaves. Hmmmm, think I’ll take on those declining reference statistics next.

Next: VIVO >