91��ý

American Libraries August 2002, pp. 62-63

By Mary Jo Lynch

On November 27, 2001, the New York Times carried a story headed “Economists Make It Official: U.S. Is in Recession”. The first paragraph of the story noted that “The group of economists that tracks business cycles made official today what has been apparent to laid-off workers and struggling businesses for months: the longest economic expansion on record gave way earlier this year to the first recession in a decade….”

The downturn, which began in March 2001, was also apparent to public librarians who noticed that circulation was increasing, while their budgets were being cut. Those librarians began calling 91��ý, asking for evidence of what Stephen E. James once called “the Librarians’ Axiom” that “public libraries prosper whenever the country is experiencing economic stringency.”

According to a 1986 article by James in Public Library Quarterly, the relationship between library use and economic conditions had been discussed for over one hundred years. He notes that one of the first references to the linkage is a statement by William Poole in the 1880 Annual Report of the Chicago Public Library and mentions a later reference to the same idea in Bernard Berelson’s classic 1949 volume on The Library’s Public. James asserts that there is “ample evidence” from the time of the Great Depression to substantiate the linkage between business cycles and public library use. But his own research, a study of economic conditions and library use in twenty large cities from 1960-1979, did not establish that “the Librarians’ Axiom” is true. According to James “overall the investigation suggests that when one uses the most rigorous statistical standards no relationship can be shown between local economic conditions and the use of public libraries.”

The article by James was the only literature 91��ý could suggest to those who asked questions on this matter. As those questions increased and National Library Week 2002 approached, the Office for Research and Statistics and the Public Information Office decided that the time was right for another study. Given budget constraints and the need to finish the work in early April 2002, we could not do anything as elaborate as James had done. Instead, we worked with staff at the Library Research Center (LRC) of the University of Illinois Graduate School of Library & Information Science to design a small study that would take a contemporary look at an old belief.

Because the LRC manages the Public Library Data Service under contract to the Public Library Association, the staff at LRC is in touch with the people responsible for statistics at many public libraries. LRC staff contacted those people at the twenty-five public libraries in the U.S. serving populations of 1 million or more and asked them to provide monthly data on circulation and visits for the last five years. Twenty-three of the twenty-five agreed to cooperate and sent data. The visits data were not robust enough for statistical analysis. However, circulation data from eighteen libraries were exactly what was needed.

Using that data and the standard methodology of time series regression analysis, LRC found that circulation has increased significantly in all the months since March 2001, when the National Bureau of Economic Research pegged the beginning of the latest recession. To do that analysis, LRC first combined the monthly circulation data from all 18 libraries and computed the average for each month. Those averages were plotted on a graph for the 48 months between January 1997 and December 2000. A trend line was established and extrapolated into the 12 months of 2001. Actual averages for 2001 are well above the trend line as shown on Figure 1.

But additional analysis makes this conclusion ever more impressive. Statisticians have determined that variation of data values in any time series is the result of four types of change:

1. Normal growth or decline over a long period of time
2. Seasonal variations
3. .Cyclical movement in the economy
4. Residual or Random factors

Mathematical formulas have been established to remove the effects of the first two types of change. LRC applied those formulas to the data and developed Figure 2, which shows cyclical variation alone (plus possible random variation). Circulation is 8% above trend in March 2001, the date when the recession officially began. It stayed well above trends, an average of 9.1% above, for the rest of the year. Technical details of that analysis are explained in a report from LRC posted at: http://www.ala.org/ala/ors/reports/Economichardtimestechnicalreport.pdf.

Does this prove that “the Librarians’ Axiom” is true? It certainly seems true for those 18 libraries in this period of economic stringency. It would take much more research to establish that those results apply to other libraries and to other points in time. But the more interesting question is “Why?” A couple of obvious possibilities come to mind: people who are unemployed check out books to help themselves qualify for new jobs and people with less money to spend get books at the library rather than buy them. To really answer the “Why?” question we need to know a lot more than is known now about how and why people use public libraries. In any case, there is some satisfaction in having statistics to document what librarians have accepted as true for over a century.