Establish Uniform Definitions
Statutory Responsibilities for Reimbursement of Nonresident Library Usage
Categories of Residence to Be Used for Data Gathering
Methods of Data Collection
The Special Challenges of Surveys and Survey Methodology
Many Wisconsin libraries already collect accurate nonresident usage statistics and use those statistics to support efforts to obtain equitable funding of nonresident use. For those libraries not yet collecting accurate nonresident usage statistics, there are at least four important reasons to begin collecting these statistics soon:
- Certain nonresident usage statistics are required for the 1998 library annual report which must be provided to the state and to each library's governing municipality in early 1999.
- County nonresident usage statistics are required for calendar year 1999 in order to ensure the required "70 percent" county reimbursement beginning in 2001.
- Certain nonresident usage statistics are needed in order for county library plans to meet the new statutory county library planning requirements.
- Good local library planning and management require accurate user statistics.
The DLCL is recommending that library systems work with their member libraries and counties soon on procedures for the collection of adequate nonresident usage statistics. Local library policies and procedures should be established for collecting the library usage statistics needed to receive county reimbursement. It is important that these policies be uniform countywide and ideally systemwide and even statewide. We hope the definitions and guidelines below help achieve the goal of accurate statewide nonresident usage statistics.
For the purpose of the new county funding requirements, it is essential that data be collected in a manner which can withstand scrutiny by your county. If your library has rigorous procedures for confirming the place of residence of each user, and if your library keeps an actual count (as opposed to sampling) of annual usage broken down by place of residence, your county will probably have no problems accepting your data. If, however, your library uses sampling to determine annual county nonresident usage, we recommend that your data collection methodology be validated by a reliable source. If you will be updating your county library plan, it is recommended that the plan address the issue of uniform library usage data collection procedures.
By statute, a county clerk may now have access, upon request, to all books and records used to determine both the annual library material loans to county residents who do not maintain a public library and the total annual library material loans. The identities of individuals using the library, however, are protected under Wis. Stat. section 43.30.
Many counties and systems in Wisconsin are currently engaged in the monitoring of local public library use patterns. This requires common agreement on how to define "public library use." The most commonly adopted measure of public library use is circulation. Circulation is also the measure of nonresident use for which Wisconsin counties will be required to provide reimbursement of at least 70 percent of cost beginning in 2001, per Wis. Stat. section 43.12. Admittedly, circulation of materials represents only one aspect of the range of services offered by public libraries. Libraries also offer a wide range of other services to residents and nonresidents alike: reference service, programming, in-house use of material, Internet use, and so on. Circulation is nevertheless accepted as a reasonable substitute for the total service provided for reasons of accuracy and simplicity:
- Circulation is generally considered to be more accurate than other measures of use. Libraries have more experience gathering this data, and there is greater consistency from library to library with regard to taking this measure than there is with regard to other measures.
- Patron use of the full range of library services generally correlates with circulation. If circulation statistics indicate that a particular category of library users account for 25 percent of total circulation, it is generally accepted that that group would also account for approximately 25 percent of usage of the full range of library services.
- A costing model that focuses on one measure—circulation—is simpler, easier, and less costly to administer than one that factors in a wider range of library services.
Today, circulation statistics are generally accepted as a reasonable representation of the total service provided by libraries, and the total operating costs divided by total circulation is generally accepted as a reasonable estimate of the overall unit cost for library use.
Establish Uniform Definitions
The purpose of library use statistics is to determine the use made of a public library by persons from the home community and those residing in taxing districts outside the primary taxing district of the public library. To complete this task it will be necessary to learn the residency of nonresident users and to measure the circulation to each identified group. Adopting the following definitions will better ensure comparability of collected data within the county, within the system, and statewide:
- Public library use. The circulation of a legally established public library's materials to users of that public library. Although public libraries provide a wide range of services, for practical reasons of data collection, circulation will be considered a valid estimate of total library use.
- Circulation transaction. The act of lending an item from the library's collection for use, generally (although not always) outside the library. This activity includes checking material out either manually or electronically, and also renewing the loan of materials previously borrowed. Each of these is reported as a circulation transaction. Items included are those circulated from all library units (such as main library, branches, bookmobiles, and book-by-mail programs) administered by the library board. Interlibrary loan items provided to the library and then circulated by the library to a borrower should be included, but interlibrary loan items sent to other libraries are not to be reported as circulation transactions. "Bulk loans," such as the placement of collections of materials in a nursing home or preschool, are not to be included. Circulation of a multi-part item representing a single title (such as a books-on-tape title containing more than one cassette) should be counted as a single circulation transaction.
- Library taxing unit. A municipality that has established a public library under Wis. Stat. section 43.52; a group of two or more municipalities or a county and one or more municipalities that have established a joint library under Wis. Stat. section 43.53; or a county taxing for a consolidated county library under Wis. Stat. section 43.57(1). Residents of a county who do not live in any existing local library service jurisdiction are not residents of a library municipality.
- Residency of patron/borrower. Patrons' residency must be properly identified and classified in order for a library to tally resident and nonresident use. In most instances, classifying residency is not complicated and is based on identifying the taxing jurisdiction to which the patron pays taxes to support library service. Residents of a municipal or joint library are considered residents of that particular library taxing unit. Residents of a town that is not part of a joint library but instead makes annual "contributions" to a municipal library are not residents of a library municipality; rather, they are considered county residents of municipalities which do not maintain a library. Likewise, residents of a county who do not live in any existing local library service jurisdiction are considered county residents of a municipality which does not maintain a library.
- Resident borrower. A resident borrower is an individual who resides within the library's legal service jurisdiction. Each library will define resident borrowers according to local policy. Some libraries may consider as a "resident borrower" an individual who owns property in the jurisdiction of the library taxing unit, even though the individual may maintain a primary residence elsewhere, arguing that the individual, as a local property owner, is supporting the library through taxes. For purposes of usage statistics, such individuals may be considered resident borrowers, if this is the policy adopted by the local library board.
Statutory Responsibilities for Reimbursement of Nonresident Library Usage
Wisconsin Statutes determine the units of government responsible for reimbursement of library use by nonresidents. Following are the statutory reimbursement responsibilities for different categories of nonresident usage:
- Usage by borrowers who reside in the same county but not in any municipal library service jurisdiction. Financial support for serving these residents must, by Wis. Stat. section 43.12, be provided by the county at a level of at least seventy percent of operating cost for this service (not including federal expenditures or capital expenditures) beginning in 2001, based on 1999 data. (Counties with a consolidated county library or a population over 500,000 are exempt from Wis. Stat. section 43.12.)
- Usage by borrowers who reside in another municipality maintaining a library that belongs to the same system. This represents use by residents who have their own "home" library either inside or outside the county but choose to borrow from other system libraries. Under the terms of system participation, member libraries must guarantee open access to all other system residents (Wis. Stat. section 43.15(4)(c)(4)). These usage statistics should be reported to the respective "home" library communities and to the system. Some county library plans and system plans provide for reimbursement of this usage.
- Usage by borrowers who are residents of other counties in your system who live in municipalities which do not maintain a library. Under system membership requirements, open access must be guaranteed to these individuals. Support for serving these residents must, by Wis. Stat. section 43.11, be provided through each respective "home" county's plan for library service. These usage statistics should be reported to the respective "home" county and to the system.
- Usage by borrowers who are residents of other Wisconsin counties not in your system. This group represents users from counties not in your system, some whom have their own "home" public library and some who do not. Reimbursement for service to residents of adjacent counties who reside in municipalities without a public library must be provided by the county at a level of at least seventy percent of the library's operating cost for this service, according to Wis. Stat. section 43.12 (counties with a consolidated county library or a population over 500,000 are exempt). Under system membership requirements, a library has the options of reducing service to or refusing to serve nonsystem borrowers in an adjacent system if adequate reimbursement was not received in the last year such cards were honored (Wis. Stat. section 43.17(11)(b)). Some systems maintain reciprocal borrowing agreements with other systems, counties, and/or libraries that provide for open access and in some cases reimbursement for cross-system-border usage. These usage statistics should be reported to the respective system, county, and/or "home" library.
- Usage by borrowers who are residents of other states. The statutes do not address public library service across state borders. Libraries may establish their own policies concerning usage by residents of other states.
Categories of Residence to Be Used for Data Gathering
Participants in this data-gathering effort must agree upon the specificity of information needed. Consistency of the data-gathering categories will improve the usefulness of the data.
- At the very least, within the home county, borrowers should be tracked according to the library taxing jurisdiction in which they reside. This would include residency categories for patrons who reside in each of the municipalities in the county that support a public library, plus a category for all residents in the outlying county area.
- Home county borrowers may be further classified according to their town of residence. This means a more complex data gathering strategy than is required, but produces a clearer understanding of geographic use patterns within the county.
- For other counties within your system, the annual report to the municipality and state will require, beginning with the report for 1998, that usage statistics be reported for each county, broken down according to users who are residents of library communities and users who are not. These statistics also need to be reported to other counties in your system so their county library plans can comply with the new Wis. Stat. section 43.11 requirement that each county library plan provide reimbursement for use of system libraries by residents of each county who do not maintain their own library.
- Borrowers who live outside your system could be identified by home county, home community, the county where they pay their library taxes, or public library system. A more detailed classification of cross-system users is necessary only for immediately adjacent counties because cross-county borrowing declines rapidly as distance increases. The annual report to the municipality and state will require, beginning with the report for 1998, that usage statistics be reported for each county adjacent to your county, broken down according to users who are residents of library communities and users who are not.
- Libraries should be aware that they may not receive reimbursement for service from any taxing jurisdiction for which they do not keep accurate usage statistics.
Methods of Data Collection
Actual usage counts are always preferable to survey estimates. All surveys, no matter how extensive and sophisticated, will have a margin of error. Almost all automated circulation systems can be configured to capture actual usage data and automatically break it down by residence category. All that is usually required is that patron records include data fields indicating the towns, villages, cities, counties, and states where significant numbers of patrons reside. Provided that patron records are accurately coded by residence, properly configured automated circulation systems can keep highly accurate and detailed records of usage by patron residence. This is a significant advantage of automated, as opposed to manual, circulation systems. All libraries should periodically ask patrons whether they still live at the address shown on their card or in the library database, in order to ensure accurate statistics by residence category. We recommend that this be done at least once a year. Annexations by a library municipality require recoding of the patron residence category for any library patrons affected by the annexation.
Up-to-date plat books and grid maps for the relevant counties help with most questions of residency. Libraries should also have on hand the most current municipal boundary maps that cities and villages produce.
Procedures for accurately coding each patron's residence category and keeping those records accurate over time should be established by each library.
(Libraries that will not be conducting manual counts or surveys can skip the remainder of this document.)
Libraries with manual circulation systems can also keep actual counts of usage by patron residence categories; however, the added staff time required will be significant. In addition, the greater the number of patron residence categories used, the greater will be the staff time required to count each separate category. Libraries that keep actual manual counts by patron residence typically use a code on the library card that indicates the residence category of each patron.
The Special Challenges of Surveys and Survey Methodology
Libraries that do not have an automated circulation system or do not conduct annual counts of usage by residence must use survey techniques to estimate nonresident usage. For data to be reasonably accurate, proper sampling and survey techniques are required.
Sampling methodologies described in Output Measures for Public Libraries (Van House, 1987) can be applied for gathering nonresident use data. Some sampling tips from Output Measures include:
- Be consistent.
- Decide in advance exactly what will be counted and how.
- Give training and guidance to all staff who collect data.
- Minimize the impact of data collection on library services and users.
- Minimize the obviousness of measurement activity whenever possible.
- Allow time for learning.
To conduct a nonresident borrowers survey, planners will need to develop specific data collection procedures, create a data collection worksheet, select a data collection or sampling period, and compile and edit the results. For consistency, all libraries conducting sampling should use uniform procedures systemwide.
Develop Data Collection Procedures
Libraries need to collect circulation information according to residency category of the library user (see above). This task will be easier for those libraries that adopt user registration methods that allow easy residence identification by category at the time of checkout. For example, non-automated libraries in Waukesha County clearly label library cards with a uniform code indicating municipal residency and regularly confirm that the cards bear the user's current address.
If current residence category information is not available from the user's library card, library staff must ascertain this information from the borrower at the time of checkout. County grid maps with street references can be used to assist the user in identifying the user's library taxing jurisdiction. Special care must be taken in ascertaining the correct taxing jurisdiction when a town and city or village have the same name.
After user residency is determined, the circulation transactions can be recorded. Data should be periodically summarized by the public library and transmitted to the survey coordinator (usually a designated system staff member).
The practical experience of the Waukesha County Federated Library System members that are not yet automated has resulted in the these general procedural observations:
Library staff are quite good at determining residency for nonresident users from nearby communities but are less accurate for unfamiliar municipalities. User residency information must be gathered accurately, since sampling periods are the basis for annual estimates which will amplify any inaccuracies. Staff should be encouraged to use plat books, grid maps, and/or municipal boundary maps for accurate identification of library taxing jurisdictions.
People often move without notifying the library of a change of address. Staff should ask patrons periodically whether they still live at the address shown on their card or in the library database. We recommend that this be done at least once a year. Certainly current address information should be verified when a card is renewed.
Renewals by telephone or computer should require verification of the users address.
Care must be taken when issuing temporary cards that the patron's residency is correct.
Some libraries issue "business" or "school" library cards. Persons using these cards can be classified according to their home address or the school or business address. A decision regarding how residency will be determined in these cases must be made before use is counted.
Public libraries collecting statistics should review the compiled period data provided by the system or the county to ensure that no typing or transcribing errors have been made.
Libraries should keep copies of all instructions provided to staff for data collection so that staff have immediate access to them if needed. New staff orientation should include a review of procedures and an explanation of the importance of the collected data.
Create a Data Collection Worksheet
In order to collect nonresident use information for each public library in the system that is manually collecting data or conducting sampling, a form of standard design should be provided to each library. Each county will require a different version of the form.
The data collection form should be divided into six broad resident classifications:
- Residents of the home community.
- Other residents of the same county. Classify residents by their library taxing unit—municipalities, joint libraries, or the county. Further classification by township is optional.
- Residents of counties in the same system. For each other county in the same system classify residents according to their specific, individual library taxing unit (individual municipalities, joint libraries, or the county); or combine all residents of library municipalities into a single category and maintain just two residency categories for each system county (individuals with a library and those without a library).
- Residents of adjacent counties outside the system. For each adjacent county in another system, classify residents according to their specific, individual library taxing unit (individual municipalities, joint libraries, or the county); or combine all residents of library municipalities into a single category and maintain just two residency categories for each adjacent county (individuals with a library and those without a library).
- All other Wisconsin borrowers. This category represents residents from nonadjacent counties in other systems. Most libraries will not wish to subdivide this category because usage from this category will typically be below one percent of total circulation.
- Out of state borrowers.
Specify the Data Collection Period
If possible, these data should be collected throughout the year so that annual statistics are as accurate as possible. For libraries that are not able to keep these statistics on a continuing basis, a sampling period should be determined. These sample periods should be uniform throughout the system and consistent for all libraries engaged in sampling. The exact dates of the sample period should be established by the public library system.
The length of the sample period should strike a balance between the interest in accurate and reliable results (which argues for a longer sampling period) and the interest in minimizing the data-gathering burden on staff (which argues for a shorter, less demanding sampling period).
One of two means of structuring the sampling period is usually employed.
- A sample can be taken on specified, scattered days throughout the year, selected by the system. No fewer than twenty individual dates should be selected, and participating libraries should be notified of these dates in advance. If the dates are randomly selected, they should, on balance, be representative of seasonal variations in service patterns. Information is collected by the libraries and transmitted immediately to the system (or the county coordinator). The system or county coordinator compiles this information for the year and makes annual estimates based on the collected data.
- Alternatively, data can be collected during specified sample weeks, determined by the system. At a minimum, four weeks should be selected—one in each season.
It is essential that the weeks chosen be "typical" or at least that they be perceived as typical. The week between Christmas and New Years, for example, is probably not a representative week for most libraries. Results are compiled by the system (or county) and applied to the entire year. A more thorough discussion of sampling considerations may be found in Output Measures for Public Libraries (Van House et al., 1987).
Compile, Edit and Audit the Results
System or county coordinators are responsible for assembling the results from the individual libraries. Where complete annual data are available, they should be used. Where sampling is done, the results should be converted to their annual equivalents to make the data comparable with annual statistics. The results of a four-week sample may be multiplied by 13 in order to translate the sample into an annual result—4,900 transactions in a four-week sample period can suggest an annual circulation rate of 63,700 (4,900 x 13).
Sample results should be examined for unexpected or anomalous findings. Data from a sample may, for example, suggest that the library's annual circulation is substantially lower or higher than is actually the case, indicating the need for an alternate way to treat the data. The results of a four-week sample of circulation may show that the library completed 4,900 transactions during the sample period. Converting that result into an annual figure suggests that the library should circulate 63,700 items per year (4,900 x 13). But, in fact, over the last five years the subject library has customarily circulated just 50,000 items per year. The result drawn from the sample might suggest that either the library's circulation has taken an unexpected 27 percent increase or the sample happened to be taken during an unusually busy time for the library.
In this case, the library or libraries involved with the nonresident survey may find it a more satisfactory reflection of the real world if the sample is used to determine the relative proportions of use among various population groups. If 490 of the 4,900 sample transactions went to nonresidents, it is more accurate for the library to state that approximately 10 percent of its total circulation goes to nonresidents, and that annually its nonresident circulation is approximately 5,000 (10 percent of 50,000 total circulation).
The coordinating system or county should also edit and occasionally audit the result to ensure the correct definitions and procedures are used.
Van House, Nancy, et al. Output Measures for Public Libraries: A Manual of Standardized Procedures. Chicago: American Library Association, 1987.
Let me know if you have any questions.
Denise Anton Wright
Public Library Administration Consultant
WI Dept. of Public Instruction
Division for Libraries and Technology
125 S. Webster St. Madison WI 53707-7841
(608-266-7270, fax 608-267-9207)