Driving Circulation Statistics: A Correlation Study of Small Town Library Parking and Circulation Statistics
The objective of this research is to determine if a correlation exists between dedicated parking and circulation in small town Illinois libraries. Data on circulation, dedicated parking and collection size was collected for 32 libraries in Illinois towns with a population between 3,000 and 5,000 people. The results suggest that there is only a minor correlation between parking and circulation. However, a strong correlation was discovered between collection size and circulation.
This study explores the correlation between the availability of dedicated parking and circulation in small town Illinois libraries. With state budgets declining and libraries under increasing pressure to justify allocation of public funds and account for expended funds, public libraries are finding themselves under increasing pressure to justify each dollar spent. As Svanhild Aabo (2009) writes, "Public libraries receive a high proportion of public funds for cultural activities and they, therefore, meet demands for more accountability" (p. 312). Expansion of parking at libraries is an expensive project and as such is a subject that merits serious consideration. However, existing literature has not explored the effect of dedicated parking on library services, and the author is not aware of any empirical data demonstrating the importance of dedicated parking to libraries. As such, we sought to ascertain what, if any, impact expansion of parking has on circulation, the statistic most often utilized by libraries to demonstrate their service to the community.
The pressure that libraries now face to justify their usefulness is illustrated by a recent news story written by Anna Devlantes for Fox News Chicago (Devlantes, 2010). The article, entitled Are Libraries Necessary, or a Waste of Tax Money? charges that the $120 million dollars spent annually by the city of Chicago on its public library system to support the Chicago Public Library could be spent elsewhere, and that the Internet has made libraries largely irrelevant. That a major news outlet approves and runs a story that presents library budgets as simply excess fat to be trimmed when money is tight is an event that should set off alarms in public library administrator offices around the state. In this present era of media and pundit suspicion regarding any government spending aside from perhaps the military, when the necessity of libraries is openly debated, spending good tax money on parking expansion is a project best carefully considered.
On November 29, 2010, the Chicago Public Library announced that the city's 2011 budget did not, in fact, leave them with a deficit. Other library systems have not been so lucky, however. In California, for example, the city of Santa Clara's City Council recently awarded a no-bid contract to L.S.S.I., a Maryland-based private equity firm, to take control of the city's library. Frank Pezzanite, the chief operation executive of L.S.S.I., has promised to cut wages and benefits in the name of a healthy bottom line. His firm currently operates 63 public libraries, and his statement on taking over the Santa Clara Branch bodes ill for its employees:
There's this American flag, apple pie thing about libraries. Somehow they have been put in the category of a sacred organization. A lot of libraries are atrocious. Their policies are all about job security. That's why the profession is nervous about us. You can go to a library for 35 years and never have to do anything and then have your retirement. We're not running our company that way. You come to us, you're going to have to work. (cited in Streitfield, 2010, p. A1).
Whether in Chicago or Santa Clara, it is a fact that libraries are coping with demands to prove that they deserve public funding in an era of massive budget shortfalls. Aabo (2009) writes that demands such as cost justification and return on investment "have been strengthening due to increasing economic pressure. There is no doubt that the pressure will increase considerably now" (p. 312). Although librarians understand the value of libraries to society, and can and do advocate eloquently and passionately for them, many of these benefits are difficult or impossible to affix with a dollar value. A clearer and simpler way for a public library to argue its case can be to stick with quantifiable measures such as circulation and number of visitors. Fialkoff (2002) points out that circulation statistics can be a potent tool to lobby for additional funds. Circulation statistics "are regularly trotted out to show the effectiveness of a library," and circulation data and budgets are often analyzed together to assess a library's performance (p. 1). Similarly, parking statistics can provide a quantitative measure of library use. However, as a single parking space can cost $950 annually in upkeep and repairs (Litman, 2011, p. 11), it would be helpful to know precisely what kind of return on investment a library might expect for such an expense.
With all of this in mind, we set out to answer a simple research question: Is there a positive correlation between dedicated parking and library circulation at small town libraries? Whatever the findings, we believed this would be of value to small town libraries' maintenance and development. If we discovered that there was indeed a positive correlation, then we might discover something close to a magic number or point of diminishing return. If we found no correlation at all, then directors of small town libraries worried about parking could at least stop doing so, and divert finite funds elsewhere. To this end, we also decided to examine circulation in light of library square footage and collection size, in hopes of being able to supply a recommendation for spending priorities. Parking is an expensive resource that requires valuable real estate, and saved money is just as good as added income. From an academic point of view, there is no existing research on this topic. What little literature that involves libraries and parking does not mention circulation. Thus, this study should be of interest both to public library employees and administrators and academics engaged in the study of library services. While further research is required to confirm results for libraries in larger communities, findings should be of interest to other types of libraries as well.
As Voith (1998) points out, parking should be viewed as what the business world calls derived demand - a resource sought as a means to accomplish a goal such as work, recreation or consumption. In regards to small town public libraries, the connection to circulation is clear: Circulation is a function of patron demand, and parking is a derived demand of that same goal. Parking is a means toward the patron's end of obtaining whatever library service he or she seeks. No studies have examined circulation as an evaluative measure of the adequacy of parking. Circulation has traditionally been used to evaluate patron interest and aid in allocation of funds for collection development. Circulation statistics have not been examined to determine adequacy of parking. The existing research on both parking and circulation data provides an overview of how these topics are important to libraries, and reveals a definite absence of research in this area.
Research in the 1960s (Hackman & Martin, 1969) demonstrated that increased parking did increase sales for shops in suburban malls (Lan & Kanafani, 1993, p. 291). Models were created with the theory that the square footage of the establishment determined the amount of dedicated parking (Shoup, 1999, p. 557). These models, which left the majority of spaces empty for most of the year, were created specifically to satisfy peak parking needs during the holiday shopping season (Arbatskaya, Mukhopadhaya, & Rasmusen, 2004). Then these same models were applied to urban and small towns without sufficient research (Urban Land Institute (ULI), 2010). The impact of the application of these models can be seen in one of the few articles related directly to libraries and parking, written in 1978, which recommended that for every square foot of library space there should be 1.5 square feet of parking (Galvin, 1978, p. 2311). Other studies involving optimization of parking focused on identifying a minimum number of parking spaces per square foot (e.g., Shoup, 1999). The prevailing wisdom has held that the ideal parking structure should be 85% to 95% full at peak occupancy, allowing potential customers to continue to find spots, without wasting valuable resources (ULI, 2010, p. 11).
Starting in the late 1980s and continuing into the present, theories and models about parking have changed. Many scholars, beginning with Donald Shoup, began to examine the effect of free parking on the costs of consumer goods, arguing that all parking should be paid in order to lower the burden on business owners (Shoup, 1999). Around the same time, researchers began to explore parking and its effects on consumer use and spending, including the effect of free parking on consumer motivation (Bacon, 1993). Waerden, Border and Timmerans (1998) more fully articulated the idea of optimum spacing, arguing that both too few and too many available parking spots will deter customers. When there are too few spots, customers do not believe they will easily find a spot, and when there are too many spaces or the parking lot is too large, customers do not want to walk long distances to their cars when carrying heavy goods and believe that a large amount of empty spaces may indicate an establishment's poor quality.
These changes in the prevailing parking theories have encouraged many communities to stop thinking in terms of classical models of minimum parking requirements. Many cities and counties are beginning to plan their parking models around the idea of the maximum number of parking spots (Hess, 2001). That is, after an ideal maximum number of parking spots in an area is determined efficient parking is promoted so that potential customers/patrons encounter 85% to 95% occupancy at peak times. Other research has focused on establishing models that will encourage wider use of public transport to reduce parking demand (Merriman, 1998).
The majority of research involving parking is explicitly concerned with businesses. But is there a home for small town libraries in this body of literature about shopping malls and downtown business districts? The authors of this study contend that these are applicable to libraries as the two share in common the goal of encouraging patrons to consume an offered product. Although public libraries do not, by and large, charge for their services, public libraries' funding is partially determined by circulation, and whether money comes from the state or customers matters little to the bottom line. The few scholarly articles that concern parking and libraries do not directly address the issue of how circulation and parking are related. Hershenson (2006) focuses on how the use of libraries as early voting locations has strained the library parking supply, and Forsyth (2006) describes how libraries built into shopping centers - and allowed to use some of the ample Christmas shopping season parking spaces - are able to save money by externalizing that cost.
Although, as Day and Revill (1995) argue, circulation data is "the strongest single element…on which to base decisions" (p. 150), none of these studies examined a possible correlation between parking and circulation. Fialkoff (2002) observes that high circulation statistics can be a bargaining tool that libraries can leverage in order to gain funding, and circulation data is consistently used for collection development purposes (Adams & Noel, 2008; Day & Revill, 1995). A library with very low circulation may find its budget primed for a cut whenever public funding is tightened; considering this, it seems curious that no study has been conducted to ascertain whether or not increasing parking has a positive effect on circulation statistics.
Of course, the importance of circulation statistics extends far beyond funding issues. Circulation data has been used to assess user needs by comparing faculty and library title selection (Dinkins, 2003), and for evaluation of approval plans (Kingsley, 1996). Circulation statistics are being utilized in personnel management to effectively staff areas of the library based on peak check out times (Anbu, 2006). It is important to note, however, that circulation data is not infallible. There is no standard for what specific information is collected (Luzius, 2005; Fialkoff, 2002). Differences in item loan periods and item renewals can affect the data (Luzius, 2005). Circulation data only counts check outs, not in-house use or after check-out use (Briscoe, 1992). In the context of this study however, we should keep in mind the power of circulation statistics in demonstrating return on public investment by way of public services.
Via survey, we collected data from 32 public libraries in towns with populations of between 3,000 and 5,000 within the state of Illinois. In any study of this type it is imperative that the researchers control for as much variance in the population variables as possible. Therefore, the researchers controlled for other available forms of transportation by sampling only small towns, since small towns typically do not have established public transportation networks that would provide a viable alternative to private transportation and the resulting need for parking facilities. Second, we decided to limit our sample to libraries in Illinois in order to control for possible state-to-state variances in user behavior. After compiling a list of every town in the state with a population of between 3,000 and 5,000, many were found not to contain independent libraries, particularly in the suburbs of Chicago where the small suburban communities rely on one large regional library. Such towns were not included in the sample. The final list consisted of 75 suitable libraries which were then polled with a data collection instrument consisting of four questions:
We sought to ascertain collection size and library size in addition to parking in order to conduct a factor analysis of the results. Population data was readily available from the U.S. Census. Examining circulation in the context of these independent variables would provide a fuller picture of the dynamics at work in respondent libraries. Because part of the justification for this project is the expense of building and maintaining parking spaces, it is prudent to learn whether other factors influence circulation statistics more than parking does. If such other correlations exist, we may then make a recommendation that funding being considered for parking expansion should be instead, under certain circumstances, diverted elsewhere.
After a list was compiled of all libraries in towns with populations of between 3,000 and 5,000, data collection was undertaken in several stages. Survey data was, in a few cases, collected via the web site Library Technology Guides, which lists yearly circulation data for public libraries and via Google maps for manually counting the number of parking spaces in a library's parking lot. After finding that email surveys often returned incomplete data, surveys were administered over the phone. Of the 32 libraries for which we collected complete data, 17 served towns with populations of between 3,000 and 4,000 and 15 libraries served towns of between 4,000 and 5,000. As part of the IRB approval process completed for this project, an assurance was made to each library that only the data would be used in the study, not the name of the library itself or the individual respondent.
The respondent libraries showed a wide variability in the levels of both dedicated parking and circulation (see Appendix). Two libraries had no dedicated parking at all, while three libraries offered a single space. At the other extreme, one library's parking lot had 53 dedicated parking spots (see Figure 1). One third of the libraries offered between 14 and 20 parking spaces and half offered between 14 and 26. The mean number of parking spots (14.32) and median (15.00) of this data set fall within this group (See Table 1). The small number of libraries offering very limited and very expansive parking resulted in a standard deviation of 12.1. Collections sizes also varied considerably, from 15,210 to 66,051. Over half of the libraries had collections of between 20,000 and 35,000 items in 2009. The mean collection size (32,205) and median (28,556) fall within this group. Circulation showed the greatest variability of all. Of all libraries sampled, the lowest total of circulated items in 2009 was 1,917, while 76,667 was the high. The mean for this dataset is 34,547 and the median 32,500, both falling into what is by far the most common circulation range: 30,000 to 39,999. Library square footage also showed significant variability, ranging from 2,400 to 11,000. Only four libraries had a square footage of 10,000 or greater, however. The mean (6,259) and median (6,000) are both indicative of the majority. See Table 1 for the mean, median and standard deviation of all variables.
The first step was to investigate the issue collinearly. In data such as these, some degree of inter-correlation between the predictive variables is to be expected. The authors decided to conduct a correlation analysis of all raw data groupings to understand how factors other than parking might be affecting circulation. As such, circulation became the dependent variable and parking, square footage, collection size and town population became independent variables. As shown in Table 2, collection size and circulation had the strongest correlational relationship of the four variables. In this table, a value of 1 is perfect correlation, and 0 would be no correlation. The highest correlation was between parking spaces and square footage. The lowest correlation was between parking and collection size. Here, it is interesting to note that square footage was not particularly correlative to collection size, which seems counterintuitive. While we could spend many pages discussing the various inter-variable relationships, for the purposes of this study it was decided that primary analysis should be devoted to determining how collection size and parking affect circulation. Once correlationarity has been established, the next challenge facing the researcher is to determine causality as such an analysis provides no such indication.
Data was analyzed using Pearson's chi-square. Although not particularly robust, this technique is particularly useful in testing the relationship between categorical variables displayed in matrix form by comparing the frequency of occurrence of actual observations to the expected relationship. In addition to the presence of categorical data, the basic requirements which the data must meet are: (a) that the frequency distribution of certain events observed in a sample are consistent with a particular theoretical distribution; (b) the events to be compared are mutually exclusive and collectively exhaustive with a total probability of 1; (c) 80% of the cells must have an expected count of 5 or more. The cell size constraint could only be met with a 2 x 2 matrix. A matrix of this size allows for a single degree of freedom since as one row and one column are compared, the remaining one is free to vary. The criteria for recoding the raw data were arbitrary but designed to produce categories of roughly equal size. Circulation was selected as the dependent variable and the data were divided into cells according to circulation of less than 30,000 and circulation of 30,000 or more (see Table 3). Availability of parking was selected as the first independent variable with the data being divided into groups of parking spaces less than 15 in number and 15 or more spaces. Finally, the size of the library's collection was selected as the second independent variable and the data were also divided into two cells, circulation less than 31,000 and circulation of 31,000 or more.
Figure 5 shows a greater grouping of high-circulating libraries to the right, among libraries with higher parking. Of the six libraries with circulation greater than 50,000, five of them have a greater number of dedicated parking spots greater than the mean. Of the 9 greater than 40,000, seven fall on the greater than side of the mean. The average circulation in 2009 for libraries with below-mean parking was 30,051 while the average for above-mean libraries was 37,880. Recall that the mean for all libraries here was 32,205. However, the relationship would be more apparent if not for three above-mean libraries (all with 15 spaces) displaying lower than average circulation figures. If we calculate only for libraries with 18 or more dedicated parking spaces, the average 2009 circulation jumps to 45,025. That is about 15% above the mean for circulation and 34% more than libraries with below mean parking. However, due to the relatively small data set, this is still insufficient to make a sweeping statement about the correlation between parking and circulation.
The first chi-square comparison was conducted with parking spaces serving as the dependent variable. As the data in Table 3 show, the comparison of the categories for parking and circulation did not produce significant results. Although a significant correlation was noted within the raw data, when the observations were divided into categories the correlation between the two variables was no longer significant. This produced an insignificant chi-square value of .318 (see Table 4). The analysis of raw and coded data do corroborate one another, which leads to the conclusion that the data set is not yet large enough to conduct a correlative analysis.
Figure 6 shows a greater grouping of high-circulating libraries at the right, among libraries with higher circulation. All but two of the libraries with a 2009 circulation of greater than 40,000 have a collection size greater than the mean. In addition, every library with a collection greater than the mean had a total circulation of over 30,000. For those with smaller than average collection sizes, the number with total circulation of over 30,000 was six out of 17. The average circulation for libraries with a below mean collection size was 25,891, while the average for above mean libraries was 46,491. This was the greatest difference in circulation among all variables, and confirms the high Pearson correlation of .772 in Table 2. The second chi-square test was conducted with collection size serving as the independent variable and these results are shown in Table 5. When the raw data were analyzed, the high degree of correlation between collection size and circulation was repeated when the data were recoded in categorical form. Table 6 reveals an extremely high correlation of .794 which produced a chi-square value of 18.295, a value that is significant at the .000 level.
Surveyed libraries fell within a population range of 3,000 to 5,000, a fairly large variance. The library at the high end of this scale served a population of 4,845, nearly 38% greater than the library at the other end of the spectrum, which served a population of 3,013. Circulation was therefore examined according to town population to determine if this was skewing data. The mean (4,202) and median (4,261) for this data set were very close. The average circulation for libraries serving below-mean populations was 32,630. The average circulation for libraries serving above-mean populations was 37,584, a difference of 13.2%. The difference was less that between above and below-mean parking and above and below-mean circulation.
Of all factors, the most significant correlative to circulation was found to be collection size. While the average circulation for libraries with above mean parking was 21.7% greater than those with below mean parking, the average for libraries with above mean collection size was 44.4% greater than those with collection size below the mean. The authors of this study therefore must make the following recommendation about parking: expansion, as an expensive project, should only be undertaken if patron demand for it is explicit. It does not, at the present time, seem to guarantee a corresponding jump in circulation, though there is enough of an indication of that in the raw data to warrant expansion of this study. At present, our data indicate that collection may be a better investment for libraries. In short, patrons may be more attracted to a library if it has a good collection than if it has an abundance of parking to allow access to that collection.
As noted above, three low-circulating libraries seemed to slightly skew the analysis of libraries by parking space. Two of those three libraries had collection sizes well below the mean (18,000 and 23,500). Thus, while it was found that parking and circulation do not have a strong correlational relationship, we cannot definitively rule out such a connection without further study and larger sample sizes. When the raw data were analyzed, a highly significant correlation (.001) was noted between these two variables and this seemed to be confirmed by the data in Figure 5, indicating that the division of circulation into greater or less than 30,000 is insufficient to reveal the relationship.
The possibility for replicative studies here is rich, as the current study focused on a single state and within a narrow population range. It appears that no large scale inquiry has yet been conducted into the myriad variables which might affect circulation in small-town libraries. With their limited budgets, this information would likely be of great help to such libraries, and as such a larger research project could have a large and positive impact on rural library services across the country. An expansion of the study into neighboring Midwestern states, and of the data set to approximately 100 libraries, is recommended. These additional observations will allow for an analysis of the data on more than the two levels present in this study. The highly significant correlation between collection size and circulation that was noted when the data were categorized makes intuitive sense, but further research would be required to determine causality. The respondent libraries should be interviewed to determine a narrative history of the library's parking and collection development, i.e. was a library forced to build more parking because of consistently high patron demand, or was the size of the parking lot part of the original plan? This will necessitate the collection and creation of trend data to determine whether collection size was necessitated by consistently high circulation, or, possibly, if a larger collection tends to attract more patrons.
Arbatskaya, M., Mukhopadhaya, K., & Rasmusen, E. (2007). The parking lot problem. Working Papers 2007-04, Indiana University, Kelly School of Business, Department of Business Economics and Public Policy.
Day, M., & Revill, D. (1995). Towards the active collection: The use of circulation analyses in collection evaluation. Journal of Librarianship and Information Science (Folkestone, England), 27, 149-57.
Litman, T. (2011). Parking management: Strategies, evaluations and planning. Victoria transport policy institute. Retrieved from http://www.vtpi.org/park_man.pdf
Copyright, 2013 Library Student Journal | Contact