Historically, the four measures included in the Library Journal Index of Public Library Service (sponsored by Baker & Taylor’s Bibliostat) have been circulation, library visits, program attendance, and public Internet computer use. Now, the design of the LJ Index is beginning to evolve. The stars have finally aligned to add a fifth statistical measure to the scoring—circulation of electronic materials, or e-circ for short. Because the LJ Index is based on data collected by the Public Libraries Survey (PLS)—a federal-state cooperative project of IMLS and the state library agencies—the Index could not add new measures until PLS did.

What is E-Circ And Why did We Add It?

The definition of circulation of electronic materials (e-circ) for PLS appears clear and stable. E-circ involves ebooks and downloadable audio and video but not licensed databases or what PLS now calls electronic collections. While database use is important, it is notoriously difficult to measure owing to the variety of databases and different ways in which their usage is counted. Apart from adding a couple of sentences to underscore that e-circ is limited to ebooks and downloadable audio and video, the definition remains essentially unchanged, including the explicit exclusion of databases.

We did not add e-circ to the LJ Index in 2015 because ten states had not yet begun to collect the requisite data. This was owing to a built-in delay in the PLS; because of survey scheduling variations, it takes some states a year longer than others to report a new item. This year, with only one state (Maryland) not yet reporting e-circ data, we decided that it was time to add it to the index. Unfortunately, this resulted in excluding Maryland libraries from the LJ Index this year. We hope to see them return in 2017 and encourage all states to be as early adopters as possible of new PLS data elements.

The PLS definition of total circulation is straightforward: the total annual circulation of all library materials of all types, including renewals. Not only was that definition unchanged between 2014 and 2015, this deliberately all-inclusive definition has been unchanged for at least two decades. Since 2014, ebooks and downloadable audio and video files have been explicitly defined as “materials,” and nothing in the total circulation definition excluded electronic materials. The definition has long said “materials of all types” and “in all formats.” Consequently, e-circ should be a subset of total circulation and thus already included in that long-standing service output. But, in practice, is it?

A thorough review of the 2014 PLS data set suggests the likelihood that there are still a few public libraries out there that do not understand that e-circ should be included in total circulation. Or perhaps the electronic materials landscape is so varied from state to state that some states—or perhaps even local libraries—have a built-in advantage because of their particular arrangements. Our hope is that adding e-circ to the LJ Index will help reveal the sources of this data variability—and potential inconsistency—so that the reporting of these two figures by libraries nationwide will become more accurate and stable. Our strongly held belief is that data is improved by being used and exposed to the light of day.

Interestingly, the consensus of opinion among decision-makers about state and national public library data on library materials and their circulation is somewhat disjointed: so far, separate data have been collected about public library holdings in various physical formats (print books and physical audio and video materials, like CDs and DVDs) as well as, more recently, in digital formats (ebooks, streaming audio and video), yet when it comes to circulating items in these disparate formats, everything is combined. Since the 2013 data set—the one on which 2015 LJ Index scores were based—a separate count of e-circ has been requested.

So, not only is e-circ a newly established data element, it also raises questions about reporting practices in recent years for total circulation. Our hope is that seeing these two figures interact in the LJ Index scoring process will help to improve the quality of both elements.

The question remains: If e-circ is a subset of total circulation, should e-circ have been considered for inclusion in the LJ Index? No other PLS figures that are subsumed in existing LJ Index statistics have been used in the Index, e.g., circulation of children’s materials and attendance at children’s programs. However, we now include e-circ because it represents a new kind of public library use, which—like as-yet-unmeasured database use—no longer requires most users to visit a physical library. The assumption that the overwhelming majority of public library use is connected with in-person visits is no longer tenable.

How E-Circ Changes the LJ Index

When we first designed the LJ Index of Public Library Service in 2008, it was based on IMLS data from FY06. At that time, most public library use was in-person, building-based activity. If someone wanted something to read or needed information, they went to a library and checked out a book or a CD, video, or DVD; attended a program event; or used a library computer. Consequently, when designing the LJ Index, we expected to see fairly strong correlations between each of the four variables that contributed to the Index score.

This year’s addition of e-circ is a game-changer in more than the obvious way. Certainly, adding a fifth measure to the index design alters it in purely statistical terms; however, it also changes the index conceptually. We are no longer working on the old assumptions about most public library use being in-person and building-based. While someone could download an ebook or an audio or video file at a library, it is probably at least as likely that they would do it remotely from their home, workplace, or elsewhere. This raises an intriguing conceptual question: What relationship, if any, should we expect between output measures that are at least somewhat tied to physical use versus measures that are more virtual?

To answer this question, we conducted a correlation analysis similar to the one on which the LJ Index was based originally. Only this time, we added e-circ into the equation (see Table 1).

What we found is that, while e-circ is positively and significantly correlated with the four long-standing LJ Index measures, the strength of its correlations with those other measures is weaker, below .25. Considering that e-circ does not require a physical library visit and isn’t necessarily strongly tied to program attendance or computer use, this was not surprising. What might surprise some is that the correlation between total circulation and e-circ—which is supposed to be a subset of it—is also relatively weak by previous LJ Index standards, though the strongest of e-circ’s correlations with other measures.

If e-circ had been any measure of library use associated with physical library visits, such weak correlations might have led us to question the appropriateness of including it at all. Given what we know about e-circ’s key difference from other measures, though, we consider it an appropriate and timely addition. This is particularly so, given what we know about the foreseeable future of the LJ Index: more new output measures are coming, and, more likely than not, they, too, will focus more attention on virtual use of public libraries.

This year the addition of just one virtual output measure means that 20 percent of a library’s final LJ Index score is dependent on virtual use. With the addition of another virtual output measure next year (Wi-Fi access usage), 40 percent of a library’s LJ Index score will depend on virtual use. We hope the evolution of the LJ Index along these lines will encourage IMLS and the state library agencies to facilitate the development of other virtual output measures to reflect more fully the range of ways Americans use their public libraries today.

The addition of e-circ to the LJ Index this year undoubtedly had an impact on which libraries achieved Star Library status. Most obviously, libraries that did not report e-circ for FY14 were not even scored this year. Similar exclusions were made early in the LJ Index’s history, when late adopters took a few years to begin reporting use of public Internet computers. It is a certainty that some libraries have been enthusiastic early adopters of ebooks and streaming audio and video; those libraries gained an edge in the LJ Index scoring process as well as the recognition of Star Libraries. It is also quite likely that some libraries, perhaps some states, developed a dramatically stronger capacity for e-circ activity than others. And, of course, we must also recognize that as e-circ is still a relatively new data element (this being its second year of reporting for most libraries), it is likely that some libraries reported more generous e-circ statistics than other libraries generating comparable output. We hope that including them in the LJ Index will encourage those who report and use the data to examine their own data and that of their peers more closely. We believe this provides a powerful incentive for improving data quality.