Research Topics and Methodology

Vendor-Supplied Bibliographic Records, 2000–2007

Using vendor-supplied bibliographic records to provide access to library materials has become commonplace among libraries in North America. Survey statistics indicate that a majority of academic libraries in the United States and Canada have outsourced their cataloging function in some form (Lam 2005). Besides outsourcing, catalogers frequently run into vendor-produced records in a national bibliographic database, such as the Library of Congress Online Catalog or OCLC WorldCat. What has been libraries’ experience with vendor-supplied records? What issues have arisen in response to their presence in a shared bibliographic utility?

A search of the literature on vendor-supplied records (“vendor records” hereafter) between 2000 and 2007 yields about a dozen journal articles and some news items. The literature covers two types of vendor records: brief bibliographic records added to a national database by a vendor and free or fee-based record sets offered to a library by a vendor.

The issues discussed can be categorized into three themes: trends and impact; quality; and management of vendor records. Although vendor records serve purposes beyond cataloging (e.g., acquisitions, resource sharing), all of the papers except one treated their subjects from a cataloging perspective.

Trends and Impact

Brief Vendor Records

Vendor records began to appear in national bibliographic databases in the mid-1990s. Unlike MARC records offered by a vendor as a cataloging service to libraries, vendor records added to RLIN and OCLC union catalogs were very brief bibliographic records originally designed to advertise an item for sale by the vendor (Shedenhelm and Burk 2001).

Catalogers have held ongoing discussions on brief vendor records. In 2000, Jeffrey Beall wrote a paper alerting the cataloging community to the low quality of the brief vendor records in OCLC/RLIN and their negative effect on cataloging in academic libraries. He observed that the minimum-level records created by foreign material vendors did not follow the minimal level cataloging standards and generally did not have authorized forms for names, series, and subject headings. This created multiple issues and problems for cataloging—both at national and local levels. Beall suspected that less original cataloging was being done because the addition of brief vendor records prevented original member input. On the other hand, the need for significant enhancement of the brief vendor records and the inability of most libraries to enhance the records in OCLC/RLIN databases would lead to duplicate editing efforts (Beall 2000).

Beall’s assessment of the situation was, he admitted, “subjective” (p. 230). Nevertheless, it compelled other catalogers to investigate the issues. Are vendor records becoming prevalent? Is their existence in OCLC and RLIN affecting members’ cataloging? If so, how? Three case studies immediately followed Beall’s publication.

Laura D. Shedenhelm and Bartley A. Burk (2001) analyzed a sample of vendor records for Spanish-language materials to find the gap between brief vendor records and full-level LC/OCLC records, thus to estimate the time and labor required to upgrade the former.

Charlene Kellsey conducted two quantitative studies on trends and the impact of brief vendor records. Her first study (2001) was a “preliminary investigation” of vendor records in WorldCat. By analyzing the cataloging source data in a sample of records for Italian-language monographs, she discovered that there had been a drastic decline, between 1996 and 2000, in the number of original member records and a corresponding increase of vendor records for Italian books. Her second study (2002) expanded the subject to include vendor records for materials in other major European languages. She found that an average of 24% of records for Spanish, French, German and Italian monographs were entered into WorldCat by vendors. By comparison, the number of libraries upgrading foreign vendor records directly in WorldCat was only one third of the number of libraries contributing original records.

The findings by the above authors largely confirmed Beall’s concern: brief vendor records not only had a negative effect on cataloging department workflows, but also caused a decrease in national level cooperative cataloging. Reaffirming the importance of “cooperative upgrading” of vendor records, Kellsey (2002) called for research in local library cataloging practices to identify “barriers” to upgrading brief records in national databases (p. 110).

Such research is yet to be done. Meanwhile, academic libraries have become increasingly interested in a full cataloging service by foreign material vendors. There has been good news: Casalini Libri, an Italian-language material vendor, has improved its MARC records to include LC classification, LC subject headings, and authorized name headings. Currently, Cassalini supplies PCC (Program for Cooperative Cataloging) core-level records along with its books (Mugridge 2004, 2007); Harrassowitz, a primary vendor of German-language materials, has indicated to the library community the company’s intent to upgrade its brief records to the minimal standards set by PCC (Brisson 2005). Beyond the enhanced MARC record services to the library community, the new initiatives from Casalini and Harrassowitz suggest the possibility that foreign material vendors, with their foreign language expertise, could become PCC contributing members (Brisson 2005).

The growing interest of academic libraries in full-level vendor records for foreign language materials was driven by multiple factors such as chronic backlogs, a shrinking pool of foreign-language catalogers, a reluctance to undertake original cataloging, organizational changes, a shift of budget resources to the creation and maintenance of digital libraries, streamlining of acquisitions and cataloging workflows, and closer working relationship with vendors (Brisson 2005).

Vendor Record Sets

In the context of library outsourcing, Banerjee (2001) took a positive view of vendor records and urged the library community to take advantage of using them to maximize cataloging effectiveness. The author described the benefits of using commercial MARC record sets, and analyzed the costs associated with purchasing, loading, and quality control.

In exploring the use of vendor record sets, most libraries have experienced both pros and cons. Libraries’ experiences, however, not only vary according to the vendors from whom they purchase records, but also have been influenced by such local factors as ILS and network settings, procedures and workflow, and, in some cases, the software program(s) a library uses to acquire and manage MARC record loads. At the 26th Annual Charleston Conference, librarians representing four different academic institutions and network consortia shared their libraries’ experiments with vendor record sets (Harwell 2007).

In providing access to government documents, the Colorado State University Libraries (CSUL) considered timeliness the highest criterion. The decision to load MARCIVE’s Shipping List Service (SLS) records into the local catalog, even though these were very brief records, was guided by the belief that having new titles immediately available in the OPAC through important access points such as title, call number, and name headings, is an invaluable service for library users (Copeland and Snyder 2000).

The fast growth of web-based publishing and the proliferation of full-text databases are making it less and less possible for libraries to catalog their digital resources in a traditional way. More and more libraries see vendor records as an alternative solution to the problem of providing timely OPAC access to their purchased or subscription titles. A national survey in 2003 reported that approximately 48% of the responding libraries and consortia had purchased record sets from vendors, such as NetLibrary, Serials Solutions, EBSCO, MARCIVE, EEBO (Early English Books Online), ARTFL (American and French Research on the Treasury of the French Language) or ProQuest. The remaining 52% of the surveyed libraries listed technical reasons for not yet purchasing record sets, such as the problems with the bulk import due to local ILS and server set up (Chen et al. 2004).

Jina Choi Wakimoto (2002) recounted California State University Northridge (CSUN) Library’s successful experience with EBSCO MARC records. The main reason for purchasing record sets for EBSCO journals rather than cataloging them title-by-title was obvious to the author: “[g]iven the number of full-text aggregated titles involved, adding even minimal level records manually to an OPAC would be a monumental task” (p. 80). Weighing the benefits to her library from EBSCO record sets, the author commended other vendors, such as ProQuest and Lexis-Nexis Academic Universe, who followed EBSCO’s example of providing MARC records for the benefit of the entire library community (p. 93).

Sanchez et al. (2006) described a “high-priority project” at Texas State University-San Marcos to load NetLibrary records into the local catalog and thus “provide simple, direct access” to more than 20,000 electronic books. However, lacking prior experience with a large record set of “unknown quality,” the cataloging staff decided to conduct a thorough investigation of NetLibrary records and solve any potential problems at the “front-end.”

For electronic resources, the availability of vendor records is directly affecting libraries’ cataloging policies and practices. It has, for example, played an important role for some libraries in their choice between single and multiple record approaches and influenced their choice of the latter (Chen et al. 2004). For other libraries, vendor records provide an opportunity to reevaluate and improve their current practices in managing electronic resources. Whether a library stays with single record approach or switches to separate records, its ultimate goal is to provide the best bibliographic information to patrons. In a recently published article, Jennifer Edwards at the Massachusetts Institute of Technology and Betsy Friesen at the University of Minnesota presented two different approaches in using vendor record sets to enhance the serials record management at their library (Edwards, Friesen, and Davis 2007).

Quality of Vendor Records

The main topic of concern regarding foreign vendor records in bibliographic utilities focused on quality. The quality analysis of foreign vendor records in WorldCat conducted by Shedenhelm and Burk (2001) concluded with the following findings: among the sample records, 10.94% contained typos, 20.31% included an incorrect choice and/or form of headings, and 76.56% included descriptive errors. Another major problem was the duplication rate (26.56%) of these records with the existing records, largely caused by the errors in description. Given that vendor records will be a permanent part of the bibliographic universe, and to ensure high quality records for both national and local library catalogs, the authors made four recommendations to OCLC: 1) training vendors on bibliographic description standards; 2) making OCLC authority files available to vendors; 3) crediting libraries for reporting duplicates; 4) increasing credits to libraries for upgrading vendor records (p. 18–19).

The majority of commercial MARC records are either OCLC/LC records or records compatible with the core level standard recommended by the Cooperative Cataloging Council’s Task Group on Standards (Lam 2005). Nevertheless, librarians often wonder how their outsourced records would compare to the full-level records in their library catalog and how the former would affect the integrity of the local bibliographic and authority databases (Sanchez et al. 2006).

Banerjee (2001) warned about the specific problems with subject headings and call numbers, as well as common errors associated with record sets from vendors. These included inclusion of records for items a library does not own, missing records or duplicate records, records for an item in a different format than that received, incorrect status or location, and other problems such as typographical errors and MARC coding errors that prevent records from loading or displaying properly.

The headings in MARCIVE’s MARC records loaded into the University of Arkansas Libraries’ catalog since mid-1999 were overall of good quality. Nevertheless, a significant percentage of them contained heading errors that affected authority control (Kulczak 2004).

EBSCO records purchased by CSUN Library were created under the guidelines recommended by the PCC Policy Committee/ Standing Committee on Automation Task Group on Journals in Aggregator Databases. However, they were machine-derived from CONSER records for the print version of the journals. The “hybrid” descriptions caused multiple problems, such as errors in coverage dates in 856 fields, conflicting data in 006/007/245 fields for the electronic version with 300/362 fields that still reflect its print counterpart, and missing fields that were required by national or local standards (Wakitomo 2002).

The problems with NetLibrary records listed by Sanchez et al. (2006) range from typos and format or description errors to problems with classification, access points, or incompatibility with local ILS. Noteworthy also is the varying quality of NetLibrary records produced at different times. Overall, there has been noticeable quality improvement.

Management of Vendor Records

Management and quality control in particular was another important theme of the papers on libraries’ outsourced records. Lam (2005) asked the participants in his survey whether the quality of these records affected the quality of their local cataloging. The author defined quality of cataloging as a combination of accuracy, consistency, adequacy of access points, and timeliness. Based on this definition, a large majority of the outsourcing libraries in his survey reported that the records provided by their vendors had not affected the quality of their library’s cataloging. However, this was only achieved because these libraries had established quality control as an integral part in their procedures for the management of vendor records.

As part of the quality control measures concerning MARCIVE’s SLS records, the staff at Colorado State University Libraries designed a “loader” to help the process of finding discrepancies and resolving differences in SLS, GPO, and in-house cataloging records (Copeland and Snyder 2000).

Wakimoto (2002) detailed CSUN Library’s treatment of EBSCO records. The pre-load procedures included viewing sample records, profiling and testing the load. However, to provide immediate patron access, the library chose to load the records into the database as quickly as possible and solve any problems that arose afterwards. The procedures for post-load modifications were kept to a minimum because one of the main reasons for using vendor records was to reduce human intervention and save catalogers’ time and effort.

In setting procedures for quality control of NetLibrary records, the guiding principles at Texas State University-San Marcos were user-friendliness, clarity of OPAC display, easy access, and user convenience in searching. Quite contrary to CSUN’s approach, the library staff here opted for a systematic front-end cleanup instead of dealing with records on an “as-found” basis. To provide high-quality records in a timely fashion, the cataloging, system, and reference staff worked together to identify as many potential problems as possible and created efficient record-editing methods to address the identified issues. Tools such as MarcEdit, Microsoft Word and Excel were engaged to methodically correct and improve the records prior to the loading (Sanchez et al. 2006).

Drawing from several libraries’ experiences with purchased record sets, Banerjee (2001) raised management issues from a macro view and described techniques to identify and solve problems. Libraries must consider a number of factors before loading record sets into their catalog. “On a basic level, three conditions must be satisfied for a record load to be successful: (1) one record is available for each item acquired by the library; (2) records are only loaded for materials that the library owns; (3) there is an obvious connection to the users between the records and items they represent” (p. 57). The author also believed it was important to examine whether vendor records are in standard MARC format, which is required for consistent OPAC display; whether they contain control numbers, which are essential for automatic updating; and finally how they conform to local standards. In addition, such loads may present technical challenges, such as potential problems with bulk import and the local server, or issues with remote access.

Future Research

Brief Vendor Records

Despite the ongoing concerns about brief vendor records in shared utilities, little has been published on the local practices regarding this type of vendor records. How cataloging departments handle a minimum-level record in OCLC (Kellsey 2002) and whether having it as a “base” record helps or hinders the production of final cataloging and patron access (Shedenhelm and Burk 2001) remain topics for future investigation. As this essay was being written, there was a heated debate in AUTOCAT about level 3 (abbreviated level) vendor records in OCLC. The discussion centered on the value of these records, which are not only brief but often erroneous. The question of whether catalogers are better off starting with no record rather than a level 3 record merits research.

Full-level Records from Foreign Vendors

At the same time, more foreign material vendors are offering full cataloging services. The enhancement of Casalini’s records and the proposal by Harrassowitz to supply full-level records are two good examples. However, full-level vendor records are offered as “companions” to materials libraries order, mostly through approval plans. They would never appear in WorldCat, or they would appear only after an embargo period. It would be interesting to learn the percentage of libraries receiving full-level vendor records through approval plans; the number of libraries that continue to copy brief records; and how many upgrade them directly in WorldCat. Another related issue is the language of cataloging. It might be worth statistically tracking how many foreign vendors use English language in their MARC records for non-English materials, how many use their native language in the cataloging, and how the non-English cataloging by foreign vendors has impacted the degree to which libraries upgrade vendor records. Finally, how the varying qualities and different degrees in fullness of foreign vendor records have affected the public users is another interesting area for future research.

Management of Outsourced Record Sets

The report by Sanchez et al. (2006) not only provides a good example of how a library employs strategies and tools for the automated cleanup of vendor records, but also suggests a general need for literature on programs, software, and procedures for efficient management of large record sets. Besides quality issues, technical problems often impose major obstacles to a successful bulk import of vendor records. Case studies in this area will especially benefit libraries that are new or less experienced with bulk import. Finally, sources of vendor records are worth studying. Knowing how records were created can be very useful in developing appropriate strategies and effective tools for manipulating vendor records. In regard to outsourcing, shelf-ready records have become a reality for more and more libraries. However, little has been written about how libraries handle shelf-ready records. What specific issues are involved in the quality control of this type of vendor records?

Vendor-supplied Cataloging

While the concept of vendor-supplied cataloging is becoming widely accepted in the library community, its implications need to be more fully explored and contemplated. For example, the WorldCat Cataloging Partners program—a joint effort between OCLC and material vendors—is providing MARC cataloging service to libraries at various levels. It will be of great interest to monitor how this and similar emerging programs are shaping the future of cataloging and technical services. Questions should be raised on both philosophical and practical levels. How do vendor records help accomplish a library’s mission of serving patrons’ needs for accurate, convenient, and timely access to library materials? Does the concept of vendor records threaten the core of the library cataloging function? What is the core mission of a library’s cataloging department? How does a library decide when and how much to outsource its cataloging? If the “bottom line” is cost, are vendor records, all factors considered, “cheaper?”

The fast-changing environment for library technical services and the continuously evolving relationship among libraries, their vendors, and patrons are forcing us to rethink vendor records from a new perspective and in the larger context of a library’s mission.

Works Cited

Banerjee, Kyle. 2001. “Taking advantage of outsourcing options: using purchased record sets to maximize cataloging efficiency.” Cataloging & Classification Quarterly 32, no. 1: 55–65.

Beall, Jeffrey. 2000. “The impact of vendor records on cataloging and access in academic libraries.” Library Collections, Acquisitions, & Technical Services 24, no. 2: 229–37.

Brisson, Roger. 2005. “Vendor-supplied LC-MARC bibliographic records for German-language assessment: an updated and critical assessment.” German Studies Library Group Newsletter 36/37: 11–24.

Chen, Xiaotian, Larry Colgan, Courtney Greene, Elizabeth Lowe, and Conrad Winke. 2004. “E-resource cataloging practices: a survey of academic libraries and consortia.” Serials Librarian 47, no. 1/2: 153–79.

Copeland, Nora S., and Maryann Snyder. 2000. “MARCIVE's Shipping List Service: Colorado State University Libraries' experience.” Journal of Government Information 27, no. 5: 547–58.

Edwards, Jennifer, Betsy Friesen, and Susan Davis. 2007. “Climbing the mountain: choosing the best path for serials record management.” Serials Librarian 52, no. 3/4: 335–39.

Harwell, Julie C. 2007. “Making our MARC: purchasing periodical MARC records from vendors.” Against the Grain 19, no. 4: 59–60.

Kellsey, Charlene. 2001. “Trends in source of catalog records for European monographs 1996-2000: a preliminary study of Italian monographs.” Library Resources & Technical Services 45, no. 3: 123–26.

Kellsey, Charlene. 2002. “Cooperative cataloging, vendor records, and European language monographs.” Library Resources & Technical Services 46, no. 3: 105–10.

Kulczak, Deborah E. 2004. “MARCIVE GPO records and authority control: an evaluation of name and subject headings at the University of Arkansas Libraries.” Cataloging & Classification Quarterly 38, no. 1: 87–103.

Lam, Vinh-The. 2005. “Quality control issues in outsourcing cataloging in United States and Canadian academic libraries.” Cataloging & Classification Quarterly 40, no. 1: 101–22.

Mugridge, Rebecca L. 2004. “Report on the ALCTS Technical Services Directors of Large Research Libraries Discussion Group Meeting.” American Library Association Midwinter Meeting, San Diego, January 2004. Technical Services Quarterly 22, no. 1: 73–76.

Mugridge, Rebecca L. 2007. “Report on the ALCTS Technical Services Directors of Large Research Libraries (Big Heads) Discussion Group Meeting.” American Library Association Midwinter Meeting, San Antonio, January 2006. Technical Services Quarterly 24, no. 3: 83–85.

Online Computer Library Center. “WorldCat Cataloging Partners.” 2007. Accessed 9 July 2007. http://www.oclc.org/catalogingpartners/wccp.htm.

Sanchez, Elaine, Leslie Fatout, Aleene Howser, and Charles Vance. 2006. “Cleanup of NetLibrary cataloging records: a methodical front-end process.” Technical Services Quarterly 23, no. 4: 51–71.

Shedenhelm, Laura D., and Bartley A. Burk. 2001. “Book vendor records in the OCLC database: boon or bane?” Library Resources & Technical Services 45, no. 1: 10–19.

Wakimoto, Jina Choi. 2002. “Utilization of a set of vendor-supplied MARC records to provide access to journals in an aggregator database.” Serials Librarian 43, no. 1: 79–95.

Prepared by Aiping Chen-Gaffey, Bibliographic Services Librarian, Slippery Rock University of Pennsylvania, aiping.chen-gaffey@sru.edu.