Content area
Library literature suggests that the Internet has profoundly affected traditional academic library operations over the past several years. To validate whether this trend holds at a mid-size academic library and to verify whether building traffic can be used as a model for predicting service intensity, correlations were run on traffic and three variables - circulation, OPAC searching, and reference desk activity. The results show that there is a strong relationship between traffic and these key functions, suggesting that library traffic can be a useful tool for library management to employ in gauging the overall, in-house use of the library. This provides a mechanism for administrators to address personnel and budgetary needs as well as a means for identifying the core number of staff needed so that the library can maintain an effective level of service.
Full text
Keywords
Academic libraries, Library users, Internet, Management, Library services
Abstract
Library literature suggests that the Internet has profoundly affected traditional academic library operations over the past several years. To validate whether this trend holds at a mid-- size academic library and to verify whether building traffic can be used as a model for predicting service intensity, correlations were run on traffic and three variables -- circulation, OPAC searching, and reference desk activity. The results show that there is a strong relationship between traffic and these key functions, suggesting that library traffic can be a useful tool for library management to employ in gauging the overall, in-house use of the library. This provides a mechanism for administrators to address personnel and budgetary needs as well as a means for identifying the core number of staff needed so that the library can maintain an effective level of service.
Introduction
As the Internet juggernaut picks up speed at universities, it is also sweeping across academic libraries at an ever-quickening pace. On the university level, distance education is no longer unusual, faculty increasingly are embracing Web-enhanced course tools, and Web sites are now acceptable as research sources for various class projects. It appears that pressure is being applied to traditional library activities/services because of the availability of information over the Internet and because of the increasing number of electronic database subscriptions available to patrons. The situation continues to be in flux. However, as a result of this advancing technology and changing expectations of users, libraries must be proactive in order to remain vital/central to the university communities they serve. They must understand the actual use of the services they offer, as well as patron's unmet needs, so that they can build on the libraries' strengths, work on their weaknesses, guide their future, and determine their long-term goals. Because there are many service components in libraries, a question arises: do library managers need to consider the entire service arena to verify the health of the library, or can one service variable be used to predict the vitality of other services? To answer this question, building traffic, circulation, online public access catalog (OPAC) searching, and reference desk questions at Southeast Missouri State University's (SEMO) Kent Library, a mid-size academic library, were evaluated for varying time spans from two semesters to several years.
Literature review
Banks (2000) provides the starting point for this present study through her article that reports the strong linear relationship between building traffic and OPAC searching at SEMO for three semesters, Spring 1998, Fall 1998, and Spring 1999. During this time, outside access to the OPAC was poor, making the physical presence of patrons in the library almost necessary. Therefore, it was reasonable to expect strong dependence on the library during that period, but would this continue over time, and could traffic predict the health of other library functions such as circulation and reference desk use?
No other publications go to the heart of this issue, but two types of papers have some tangential relevance. First, many authors have written "feeling" pieces about one or another slice of a library's operation, e.g. reference, without giving any hard numbers or without specific data supporting the conclusions. These articles show the pressures reference services face because of the Internet. Two examples are papers written by Lancaster (1999) and Lipow (1999). Lancaster questions a goal of librarians "to provide access to any type of information for anyone, at any time, anywhere", and counters that the library profession needs to remember to provide warm, friendly service. Lipow (1999) discusses the downward pressure the Internet has had on reference service and suggests that if point-of-need reference service is to survive, librarians' work will need to become more visible. Second, various authors report some data collecting (Hiott, 1999), surveys (Tenopir, 1998; Basu, 1995), and formal observations with a study component (Zumalt and Pasicznyuk, 1998) on answering reference questions. As research background, these articles describe the impact of the Internet on various library operations, especially reference service, and affirm that a paradigm shift is afoot. They lend support to the idea that because of a sea change in academic library use, library administrators will need to identify what level of service is critical to maintaining the health of the library organization. The goal of this study is to aid in that effort.
Methodology
There are four sets of data considered in this study:
(1) Building traffic is the total number of people exiting the building counted by the front and back exit turnstiles. Though it is true there are some offices/facilities, e.g. William Faulkner Center, student computer lab, and Writing Center, that bring people into the building who might not be using library services, it is not possible to discern who these library visitors are. The investigator believes, however, that these numbers are relatively small, do not vary much from semester to semester, and might even represent people who come to the building for dual purposes. Moreover, these people may merely shift the constant value in a linear model rather than control the slope of the relationships.
(2) Circulation is the total number of items charged out at the circulation desk. The figures in this data set are generated each time a circulation transaction is made on the NOTIS system and include renewals because patrons must bring material back into the library for that purpose. Reserve material and interlibrary loan circulation numbers are not included in this variable.
(3) Just as NOTIS tracks circulation, it also provides a daily accounting of the number of initiated author, title, subject, call number, and keyword searches. Transaction logs also report the total number of searches keyed into the system. There are currently 19 "locked" terminals that access only the OPAC. In the reference area, there are 25 terminals that provide access to the library's OPAC, online databases, and Internet. As a result of machine upgrades this past summer, it is now easily possible to search the OPAC outside the library.
(4) Library patrons ask for various kinds of help at the reference desk, including directional, computer searching, and research questions, etc. After answering each question, including a few telephone questions, the reference librarian on duty tick marks the statistics sheet as a means of tracking usage.
Monthly statistics for each of the four variables were entered into the SigmaPlot graphics software and the data were subjected to various correlations. Not only linear correlations were attempted but data were also subjected to polynomial fits. In general, second order and higher order polynomials did not improve the fitting parameters. Therefore, only linear relationships, along with their "r" values, which signify the strength of the relationships, are presented.
Results and discussion
Traffic
Figure 1 shows that the total number of people coming into the Kent Library building has steadily declined since the beginning of 1996. This decline cannot be attributed to enrollment, for enrollment of undergraduate and graduate students has increased each fall since that year. This number includes students who are not enrolled in any on-campus classes. However, because this group made up only
8 percent of the total enrollment in 2000, their impact is minimal. Figure 2 clearly illustrates this rising enrollment and falling traffic phenomenon.
The departure of two academic departments and their classes taught in the library also does not figure into the decline in traffic from calendar year 1996 through the summer of 1998. Since the departments and classes were continuously housed in the building during that time, their residence played no role in changing the traffic pattern. Their exit in the summer 1998, however, may have had some impact on the intensity of people coming into the building. To determine the influence of their move, correlations were run on the relationship between building traffic and circulation while the classes/offices were in the library and after they had left. Figure 3 shows that the slopes of linear relationships observed between traffic and circulation for the two periods are parallel. Therefore, these data clearly show that circulation patterns did not change though the shift in the x-coordinate indicates certain patrons who came to the library did not utilize circulation.
Even though faculty offices and classes moved out of the library, the reserve collection continued to be used by a host of students across various disciplines. To determine if reserve use, as a gauge for library assignments by faculty, factored into the decline of traffic, the author examined reserve use statistics for each of four fall semesters. During this time, e-reserves were not available. A mixed picture emerged, going from a low in fall 1996, to a high in fall 1997, to a low in fall 1998, to a further low in fall 1999. The lack of consistency of reserve collection use suggests that faculty have not gradually stopped assigning reserve readings and that a diminution of reserve reading assignments has not necessarily contributed to a decline in traffic.
Though there was no clear trend to the number of reserve readings assigned by faculty, there was a definite increase in the number of students coming into the library because of bibliographic instruction classes. The number increased from 3,233 in 1998 to 4,524 in 1999. The number of requests for bibliographic instruction classes increased 57 percent in 1999 from the previous year. Even with this increase, the building traffic overall continued to decline.
Kent Library has adopted several key strategies to deal with declining traffic, including the purchase of new material, the addition of several new electronic reference databases and increased numbers of Internet-- connected terminals in the reference area that permit e-mail and Internet access. In addition, over the last half of the 1990s, there has been an increase in the number of terminals available to students campus-wide that allow them to search not only the electronic reference databases without having to come into the library but also sites on the Internet. It seems reasonable then to suggest that the Internet has clearly been the prime reason for the decline in building traffic because of the convenience of terminals spread across campus, the availability of reference databases from these terminals, the acceptance of using Internet resources for class assignments, and the sheer volume of information available on the Internet.
Circulation
Circulation has also dramatically declined at Kent Library (see Figure 4). The downward trend cannot be explained by enrollment because, as suggested earlier, enrollment has gone up each fall for the last four years while circulation has gone down. It also cannot be explained by the fact that the library is not buying what students want/need since they have been able to request specific titles since 1998. University faculty historically ordered needed material through a departmental allocation system. Since the fall, 1999, faculty has participated in the development of an approval plan process. Instead, one only needs to look at the drop in building traffic to see why there is a drop in circulating material. Figure 5 ties these two variables together, strongly indicating linear dependence between circulation and traffic. The coefficient of linear regression, i.e. r has a value of 0.988, means that the two move in lock step with each other. Thus, when fewer people come into the building, there are fewer items being checked out. Conversely, if the traffic were to increase, circulation would increase. Consequently, libraries should develop
strategies of bringing people into their doors to maintain their vitality.
Like SEMO, the Association of Research Libraries (ARL) reported a downward slide from 1995 until 1998 in circulation statistics. However, the data listed on their Web site (2000) show an increase of 1 percent for 1999. They also listed a 1 percent increase in the total number of students enrolled. No such trends were revealed from our analysis at SEMO. It would be interesting to know the reason for the upsurge in circulation last year at these ARL libraries. Whether this enhanced circulation use will persist and whether there is a corresponding percentage increase relative to the total number of students will be two points to watch in the future. The Web site does caution that different libraries lend books for different lengths of time, and this indeed would factor into the number of circulation transactions a library might experience.
OPAC
Transaction logs began providing some information about OPAC searching at Kent Library two months after NOTIS became available in January 1988. Each day a report indicates the total number of searches initiated as well as the number of author, title, subject, call number, and keyword searches keyed into the system. Unlike the logs at many libraries, these have never shown the actual search or the hourly rate. However, they do clearly illuminate which types of searches are the most popular and the depth of interest in the library's current collections. They also provide raw data that provide the means to monitor OPAC use over time. At SEMO, for example, OPAC searches went from 99,515 in fall 1998 to 84,579 in fall 1999, a 15 percent decline. It is true that a comparison of only two semesters does not validate a trend, but the drop in OPAC searching makes sense since traffic into the building had been decreasing. Indeed, this relationship between OPAC searching intensity and building traffic numbers is confirmed when a correlation is run (see Figure 6). The graph highlights the fact that OPAC searching is dependent on building traffic because of the fairly strong positive correlation (r = 0.81) between the two variables. Therefore, when traffic declines so does the level of OPAC searching.
OPAC searching also ties into circulation intensity, as has been depicted in Figure 7. As the level of OPAC searching increases so does the level of circulation. In 1999 at Kent Library, the r-value for this relationship was 0.90, a strong correlation. This finding is not surprising. Since traffic influences the intensity of OPAC searching, traffic is a key factor to both the use of the online catalog and circulation.
Reference transactions
Reference transactions, i.e. patrons asking various types of questions at the reference desk, also experienced a downturn from calendar year 1998 to 1999. In 1998, there were a total of 12,321 reference transactions, while in 1999 there were 11,869. These numbers are somewhat surprising because Kent Library purchased additional public access terminals for the reference area, acquired additional online databases, provided more bibliographic instruction classes, increased the number of hours the reference desk was staffed, increased manpower, developed instructional materials, and continued new faculty orientation, suggesting that there would be greater librarian intervention and/or interaction. What is not surprising is that the relationship between traffic and reference desk activity is similar to the relationships between traffic and circulation and traffic and OPAC searching at Kent Library.
There is strong dependence between the amount of traffic coming into the library building and the level of reference activity (see Figure 8). With an r-value of 0.95, it is clear that the level of traffic influences the amount of activity going on at the reference desk.
Conclusions
This study shows that there has been a dramatic decrease in building traffic, circulation, OPAC searching, and reference transactions at Kent Library over time. It also shows that circulation, OPAC searching, and reference transactions have a strong dependence on building traffic. Thus, as traffic goes so go the other three operations. This is the most important finding of the study. The rate of building traffic controls the rate of three other traditional library activities at Kent Library.
The research suggests:
Building traffic can be used as a mechanism for monitoring the level of circulation activity, OPAC searching, and reference service. Because the relationships are strong, going from r = 0.9998 (which is almost a perfect correlation) to r = 0.81, library administrators need only look at the building traffic data to know the level of use of the other areas. This provides an avenue for identifying personnel needs at circulation and reference and for considering strategies for making sure that there is appropriate help for library patrons who use the OPAC.
Not only are circulation, OPAC searching, and reference desk questions dependent on traffic, but they also have a strong link with each other. For example, the correlation between OPAC searching and circulation has an r-value of 0.90. Thus, there is a definite relationship between how many OPAC searches are initiated and the number of circulated items.
Many librarians do not acknowledge the pressure the Internet is placing on traditional public services.
Their response is that we are providing the academic community with what it wants, and high use of the online databases must outweigh any downturn in traffic, circulation, reference desk use, and OPAC statistics. It is true the Internet is a wonderful tool that must be harnessed for its educational/scholarly/teaching power. However, if fewer and fewer people come into the library and/or if those who come mainly want to sit in front of computer screens, professional employment and the availability of library services will be adversely impacted. Personnel, morale, and funding issues are at stake. In addition, students who are not exposed to the vast array of different library materials, who do not learn various search strategies and search terms, and who do not learn how to evaluate information lose an important component of their liberal education. Though this study does not identify the optimum level of traffic necessary for the continued health of the library, it does offer a mechanism by which library administrators can monitor current service with an eye toward the future. It is of utmost importance to identify and encourage the library's core constituency to continue to come to the library so that the downward slide can be attenuated. Until then, all possible strategies must be employed to advocate the library's importance because of its dynamic learning environment and its necessary place in the life of the university community.
References
Association of Research Libraries (2000), "Trends in ARL libraries", ARL Statistics, http://www.arl.org/stats/ arlstat/1999t1.html
Banks, J. (2000), "Are transaction logs useful? A ten-year study", Journal of Southern Academic and Special Librarianship, Vol. 1 No. 3, http:/www.icaap.org/ SouthernLibrarianship/.
Basu, G. (1995), "Using Internet for reference: myths vs. realities", Computers in Libraries, Vol. 15 No. 2, pp. 38-9.
Hiott, J. (1999), "Making online use count", Library Journal, Vol. 124 No. 16, pp. 44-7.
Lancaster, F.W. (1999), "Second thoughts on the paperless society", Library Journal, Vol. 124 No. 15, pp. 48-50. Lipow, A.G. (1999), "In your face", Library Journal, Vol. 124 No. 13, pp. 50-2.
Tenopir, C. (1998), "Reference use statistics", Library Journal, Vol. 123 No. 8, pp. 32-4.
Zumalt, J.R. and Pasicznyuk. R.W. (1998), "The Internet and reference services: a real-world test of Internet utility", Reference & User Services Quarterly, Vol. 38 No. 2, pp. 165-72.
The author
Julie Banks is a Professor at Southeast Missouri State University, Missouri, USA.
About the author
Julie Banks is Professor, Kent Library, Southeast Missouri State University. E-mail: [email protected]
Copyright MCB UP Limited (MCB) 1999