The ALA executive board had expressed concern about the lack of time available for libraries to prepare for the change in series authority treatment. It stated that libraries needed that time to determine the effect of the decision and the options for providing continuing series authority control.¹⁰ To date, no formal studies have been published. The following literature review covers opinions that have been expressed about the decision’s potential consequences for libraries as well as a set of responses to an informal survey about changes to local practices following the series decision. The survey was distributed through the PCC electronic discussion list, PCCLIST, which is accessible only to PCC members and focuses on cooperative cataloging issues.

Reference librarians have commented on the adverse outcome the decision makes for providing reference service to patrons. Although stopping short of chastising the LC, Mitchell and Watstein list the many ways in which series authority control affects users, including finding series, classifying works, distinguishing between series and subseries, and supporting the incorporation of the principles of Functional Requirements for Bibliographic Records in catalogs.¹¹ They point out that in FY2004, bibliographic records created by the LC included 82,447 series statements. Donlan notes that the PCC creates more series authority records than the LC, but also points out that the organization’s members would have to nearly double output of such records to sustain the number created before the LC’s policy change.¹² She suggests that libraries may need to hire catalogers to counter the effects of the LC’s series decision.

McElfresh points out that users are hindered by keyword searching and Google-like search engines.¹³ She cites Tillett regarding the importance of a controlled vocabulary for precision in searching and as being necessary for creation of the Semantic Web.¹⁴ McElfresh therefore argues that series authority control is important and, if it is to be continued, libraries must commit to taking on the task no longer being shouldered by the LC. By sharing the responsibility of series authority work, the effect of the LC’s decision can be lessened. However, she does not attempt to assess how much work this adds to cataloging departments or how they are to finance this added burden.

In the fall of 2007, a year after the LC’s announcement, a query sent to the PCC discussion list asked list members for input on how their libraries were handling the series issue. Staff at the University of Colorado at Boulder have continued their practice of providing series authority control by creating new series authority records and editing bibliographic records with untraced series, which should have been traced.¹⁵ They have found a small increase in the number of new series authority records that they need to create, but nothing overwhelming. They found that of fifteen hundred new records, only twenty-four (less than 2 percent) had untraced series headings. They, as well as some other responders to the survey, indicated that they are looking into outsourcing their authority control work.

New York University copy catalogers search all series and flag anything new to their system for later review by serials-experienced copy catalogers.¹⁶ Original catalogers continue to create new series authority records. Their impression is that there have been relatively few new series for which they would have expected the LC to create an authority record in the past. Duke University catalogers reported that much of their authority control processes are automated.¹⁷ The effect of the LC’s series decision was small, resulting in manual corrections to five series over the course of a quarter that were deemed to be because of the LC’s suspension of series authority record creation. Northwestern University has a loader program that compares untraced series headings with authority records and automatically adjusts the series tracing as appropriate.¹⁸ If it cannot perform this action, a cataloger is notified by e-mail. The program also now compares field 490 (with a first indicator of “0”) with existing authority records. Review of e-mail messages takes only minutes per day. In the year since the LC’s authority decision, the number of series authority records needing to be created has only increased slightly. The University of Georgia, University of Florida, and Indiana University did not comment on the effect of the decision in their libraries, but did say that they continue to trace series because they consider series to be an important access point for public and other functions of their libraries.¹⁹

OSU is a land, sea, sun, and space grant institution with approximately nineteen thousand students and eighteen hundred faculty. The OSU Libraries’ holdings include more than 1.4 million volumes, 14,000 serial subscriptions, and more than 500,000 maps and government documents. A main library and veterinary medicine library on the main campus are complemented by two branch libraries serving remote facilities of the university.

The Technical Services Department includes serials and monographs cataloging units and a digital production unit in addition to acquisitions units. The three units involved in cataloging include 2.5 full-time equivalent (FTE) catalog librarians and 8.5 FTE paraprofessionals. The cataloging units participate in the Program for Cooperative Cataloging Name Authority Cooperative Program (NACO), Subject Authority Cooperative Program (SACO), and Cooperative Online Serials Program (CONSER). OSU Libraries has not yet been declared independent for creation of series records.

OSU Libraries acquires approximately fifteen thousand monographs annually in addition to receiving about five thousand government documents. Of the firm ordered and approval plan monographs, approximately 85 percent have cataloging copy (contributed either by the LC or a member library) available through the OCLC bibliographic database with full-level cataloging, including call numbers and subject headings. Most cataloging copy is not scrutinized for authority control and is downloaded in a “fast-cat” process by a lower-level paraprofessional. The remaining 15 percent require subject analysis or original cataloging or both. Original cataloging is done by a monographs cataloger and upper-level paraprofessionals; series encountered during original cataloging are searched in the LCNAF and traced according to the series authority record or, if no authority record is found, according to Anglo-American Cataloging Rules, 2nd ed. (AACR2) and Library of Congress Rule Interpretations (LCRI).²⁰

Between June 1, 2006, and December 31, 2007, new bibliographic records downloaded into the OSU Libraries’ catalog were reviewed periodically for the presence of untraced series statements (MARC field 490 with a second indicator of “0”). This was done using the “create lists” function in the OSU Libraries’ Innovative Interfaces integrated library system. The list was then sorted alphabetically by series title for ease of review. The printed reports listed the source of the catalog record, title proper, series statement, and publication date.

Each series title was then searched against the LCNAF, which is accessible through OCLC Connexion. If the form of the series title on the authority record matched the form in the series statement, the field on the bibliographic record was revised to a series statement/title-added entry (MARC field 440). Proposed changes to MARC 21 to accommodate a new method of recording and tracing series titles were approved by the Machine-Readable Bibliographic Information Committee (MARBI) at the ALA Annual Conference in Anaheim in June 2008. This occurred during the drafting of this paper. Since this study was conducted before these changes were approved, and since the changes have not yet been put into effect, this paper reflects longstanding practices of series title transcription and tracing.

If the form of the series title in the LCNAF did not match the form in the bibliographic record, the second indicator was changed to “1” and a field 830 (series added entry—uniform title) added to the record with the authorized form of the series title.

If no LCNAF record was found for the series title, the library’s catalog was checked to see if records with the series title already existed in the database. If they did, the cataloger edited the record to make sure that the new record’s series tracing matched the records in the catalog. If no other records were found in the catalog with the same series title, the series was traced as it appeared on the piece (i.e., changing the field 490, first indicator “0,” to field 440, second indicator “0”) unless the title was generic. Generic titles, such as Annual report, were revised to follow AACR2 and LCRI in constructing a uniform title by changing the first indicator to “1” and adding a field 830 with the necessary qualifier.

A tally was made of the treatment each series title received. Data were then entered into a spreadsheet. For each of the eight times untraced series were reviewed during the study, the following data were determined:

• number of new records entering the system
• number of new records with untraced series
• number of untraced series statements
• number of untraced series statements by source of the bibliographic records (characterized as LC cataloging after June 2006, Government Printing Office (GPO) cataloging, or other)
• number with publication dates before 1999
• number of series titles represented by an authority record in the LCNAF and how the bibliographic record was modified (changed to field 440 or fields 490/830 or left untraced)
• number of series titles not represented by an authority record in the LCNAF and how the bibliographic record was modified (changed to field 440 or fields 490/830 or left untraced)

Analysis of data indicated that a significant proportion of untraced series were present in older bibliographic records that entered the system because of retrospective conversion projects. To determine how much of the work of reviewing untraced series was because of these projects, works with a publication date prior to 1999 were examined as a separate subset of bibliographic records. Since most new purchases had publication dates in the past three years and most retrospective work involved works with publication dates of thirty or more years ago, 1999 seemed like a logical cutoff date to divide recent from older publications.

Of the 53,911 records added to OSU Libraries’ catalog during the eighteen months of the study, 977 (2 percent) had an untraced series statement (field 490, first indicator “0”); see table 1. Of these, only 60 (6 percent) were records created by the LC after the 2006 decision was made (see table 2). The majority of records (64 percent) with untraced series statements came from the GPO. Many untraced series were found in older records (i.e., publication dates of 1999 and earlier), the result of an ongoing OSU Libraries serials retrospective conversion project as well as the cataloging of older materials in the Atomic Energy Collection in Special Collections. These older materials totaled 266 records (27 percent), although some may have overlapped with the GPO records just mentioned.

Of the 977 records with an untraced series statement, 545 (56 percent) were represented by an authority record in the LCNAF. For 96 (10 percent), the authority record was for an untraced series or a quoted note. This is the only group of series statements with authority records that were traced and coded correctly. Another 40 (4 percent) also should have been recorded as untraced series, although no authority record for the statement was found. This brings the total for untraced series titles recorded correctly to 136 (14 percent); see table 3.

For another 334 records (34 percent), the series was recorded in the same form as either the LCNAF or as it would have been if current AACR2 rules had been consulted. In other words, these series statements were coded incorrectly when current cataloging rules were consulted. These were most easily converted to traced series in the system by changing the coding because the content of the fields was already correct.

The remaining 494 records (51 percent) had the series recorded differently than appeared in the LCNAF or the way the series would be traced according to AACR2 rules. Changing these to their correct form and coding required the addition of a field 830.

Over the course of the eighteen months of this study, approximately 54 records with untraced series statements were added to the catalog monthly. The time required to research these series titles in the LCNAF and make changes to the catalog was about two hours each month.

The data answer the question as to the extent of work needed to ensure continuing authority control of series titles in a catalog at an institution such as OSU Libraries. Of the untraced series titles in the study, 57 percent had existing authority records in the LCNAF. For this proportion of records, an automated authority control service would have been able to do the authority work and thereby reduce the workload of correcting series titles. Only 25 percent of all untraced series titles differed from their authorized forms. If an integrated library system indexed MARC field 490, these would be the only ones that require correction to provide access to the authorized series titles.

An additional 27 percent would be traced differently if an authority record were created according to AACR2. For these series titles, a cataloger must either search for the correct authority record or construct a uniform title heading and record this data in MARC fields 490 and 830. In an integrated library system that indexes field 490, this 52 percent (the 25 percent of series titles in the LCNAF plus the 27 percent not in the LCNAF) represents the work needed to make these series titles accessible in an online catalog. In a system that does not index the 490 field, an additional 35 percent will require authority work and the retagging of the 490 field, for a total of 87 percent of bibliographic records with untraced series.

Even with this high percentage needing revision, relatively few records with untraced series titles are being added to the catalog. In about two hours each month, staff can identify these records, search for and download series authority records from the LCNAF, and then make any necessary changes to the bibliographic records. These procedures are easily absorbed by staff in the current workflow. In a library using automated authority control services, the time needed to complete all corrections would be considerably less.

This study identified three main sources of untraced series statements in bibliographic records: the LC, the GPO, and retrospective conversion projects that loaded older bibliographic records into the OSU Libraries catalog.

The number of untraced series statements in LC records is very small, with approximately three or four records with untraced series added each month. Many LC records may initially have had untraced series statements that were revised by other libraries before being downloaded into the OSU Libraries catalog. The statement from some catalogers that having many libraries shoulder the additional series authority work caused by the LC’s decision would soften the decision’s effect is probably correct, but this would need to be verified in a different study.

If the rate at which new series are created by major publishers remains the same as the rate during this study, catalogers should be able to keep up with series authority record creation and bibliographic maintenance. If, however, series title changes increase or the number of new series increases, series authority record creation and bibliographic maintenance would need to be increased as well. Expansion of series authority training programs would help mitigate this trend should it appear.

As the LC revises its priorities and focuses its resources on serving Congress rather than acting as the de facto national library, the degree to which it perceives its success at abandoning series authority work may lead to its abdication in other areas of library leadership.

The GPO stated in June 2006 that it would not follow the LC’s lead and would continue to create series authority records. However, the rate at which they do so must be slow, since so many of the series on GPO bibliographic records lack authority records. Also, many of the series in GPO records were coded as untraced when their authority records indicated that they should be traced. GPO does not appear to be following its own policy and is instead choosing to slow down its creation of series authority records as well as not tracing series that have existing series authority records. These practices have created the majority of records needing scrutiny and change in this study. Since OSU Libraries purchases most of its GPO records from a third-party vendor, it is looking into having the vendor authorize the series headings before sending them.

Interestingly, a significant proportion of the series authority control work currently done at OSU Libraries appears to stem from projects involving older materials. Some records for these older materials reflect earlier cataloging practices; others are minimal-level records probably created during other libraries’ retrospective conversion projects. No series authority records are present in LCNAF for many of these older materials’ series statements. Once the retrospective conversion of serials and special collections is completed, the time needed to process untraced series should decrease noticeably.

The burden of additional work predicted by some at the time of the LC’s announcement has not materialized for OSU Libraries. On the other hand, other libraries’ experiences could differ from OSU Libraries’ depending on the type of library, types of materials collected, and the degree to which those materials are published in series. For example, a research institution collecting more gray literature than OSU Libraries might see more series titles not represented in the LCNAF. Similarly, public libraries, which often collect children’s books published in series, might find that authority records for these series titles are unavailable.

This study enabled OSU Libraries catalogers to learn about untraced series titles in the bibliographic records that are added to the catalog. The LC’s decision to stop creating series authority records and to treat all series as untraced has had only a minor effect on cataloging workflow. Instead, a significant proportion of series authority work was generated by the cataloging of federal documents, retrospective conversion projects, and the addition of older materials to OSU Libraries’ collections. Government document records exhibit poor series authority control and require a significant amount of attention. Still, such authority maintenance does not appear to be a significant burden at this time. Series will continue to be tracked as records are added to the OSU Libraries’ catalog. With the baseline provided by the current study, changes in series authority control workload should be detectable in the future.

Research into the effect of the LC’s decision should be conducted at other libraries, including larger institutions and other types of libraries, to determine if the findings from the OSU Libraries study can be replicated. Repeating this research in the future could also shed light on whether series authority record creation can keep pace with the appearance of new series. Such studies would provide a larger context in which to assess the effect of the LC’s series authority decision on the cataloging community.