Cataloging production standards for non-western languages: from a project to permanent standards.

Introduction

Much has been written about the ongoing struggle to identify, implement, and maintain meaningful production standards for catalogers. (1) This paper presents the methodology used to create production standards for the Area Studies Cataloging Section (ASCS) of the Indiana University Libraries in Bloomington, Indiana. The process grew out of temporary production expectations used in a nine-month Backlog Reduction Project in the Cataloging Division, of which the ASCS is a part. When the project concluded, the section discussed permanent standards for itself.

Environment

Librarians and high-level support staff responsible for the cataloging of non-Western language monographs collected by the Indiana University Libraries Bloomington Campus comprise the ASCS. This campus has programs in a wide range of non-Western languages, including East Asian (Chinese, Japanese, and Korean, also known as CJK), Slavic and East European (i.e., Russian, Polish, Czech, Slovak, Bulgarian, Serbian, Croatian, Hungarian, Estonian, Belarusian, Ukrainian, Latvian, Romanian, Albanian, and Greek), Middle Eastern (i.e., Arabic, Hebrew, Persian, and Turkish), other Asian languages (i.e., Mongolian, Uzbek, Kazakh, Georgian, Armenian, Tibetan, Nepali, Hindi, Sanskrit, and Thai), and African (i.e., Amharic, Hausa, Bantu, Swahili, and Twi).

Although staffing has fluctuated in this section over time, typical of technical services units around the country, eleven staff members basically make up the section. There is one staff member for each of the East Asian languages of Chinese, Japanese, and Korean, although the librarian responsible for Korean is also the CJK serials cataloger. The Slavic and East European languages have two support staff members and two librarians, although one of the librarians manages the section, and the other also coordinates the NACO/PCC (Name Authority Cooperative Program of the Program for Cooperative Cataloging) for the Technical Services Department. Two support staff members and one librarian cover the Middle Eastern languages. One of the support staff members spends a lot of time on computer support for the department, while the other also catalogs the monographs in African languages. The librarian works only 25%FTE (Full Time Equivalent) in cataloging. Our eleventh support staff member catalogs the Tibetan, Sanskrit, Hindi, and several of the Asian languages of the former Soviet Union and also helps out with the cataloging of Russian. Each cataloger specializes in a language or language group that (s)he has studied or knows natively, but many must also catalog unfamiliar languages with the aid of dictionaries and grammars.

The Backlog Reduction Project

In the fall of 2001, the Cataloging Division as a whole faced a sizable backlog problem. Some of the ASCS backlogs had traditionally been large, but all of the backlogs in the division had grown over the previous year for a number of reasons. For two and a half months at the end of 2000, all cataloging in the university system stopped when the Indiana University Libraries switched their online library catalog from NOTIS to SIRSI Unicorn. Once the new system became available, catalogers took a while to adjust to the new online system, with work proceeding slowly at first. Within a few months after the implementation of the new system, cataloging of new materials again halted for several months so that retrospective conversion could be completed. Over the course of this period, various units in the division also undertook other smaller projects.

By mid-year 2001, most of the units in the Cataloging Division faced backlogs larger than the norm, with the ASCS being no exception. The Area Studies backlog totaled approximately 6,500 monographs. Many of the materials for which the ASCS is responsible are in languages that do not normally "grow" cataloging copy over time, so they require original cataloging no matter how long they linger in a backlog.

The library administration decided that the Cataloging Division would undertake a nine-month project to eliminate these cumbersome backlogs beginning October 1, 2001, and ending June 30, 2002. In order to maximize the effectiveness of the project, it established the following division-wide "rules":

* Stabilize the work: Additional projects would be avoided whenever possible for the duration of the project. The primary focus of the staff would be the elimination of the cataloging backlogs. Other than the cataloging of the new receipts identified as "rush," staff would have a single-minded focus on eliminating the backlogs.

* Make temporary reassignments: Some staff would be reassigned from their normal work to assist in areas that had larger backlogs.

* Set goals for increased productivity: Although the cataloging department did not normally have cataloging quotas or production norms in place, all participating catalogers received an expected level of work for the duration of this project.

* Request additional funds: Extra money increased the hours of part-time catalogers and student searchers working on the project.

* "Protect" positions: Positions involved in the project, if vacated, would be filled as soon as possible--not cut or held open for salary savings.

* Do what is necessary: Copy cataloging intentionally sacrificed some quality during this project. Management did not require some routine editing normally undertaken, and encouraged catalogers to do only the necessary editing to make a record accessible to patrons while not undertaking all of the editing they would ordinarily do. For example, they could dispense with the addition of a 504 field to reflect the presence of a bibliography.

These "rules" were all straightforward, except for setting the production goals for catalogers who were not used to them. Introducing the concept at a time when productivity was expected to be higher than normal made it especially stressful on them. Since not all cataloging is equal, among the variables that had to be considered when setting the monthly production goal for the ASCS staff were:

* Copy vs. original: If a cataloger worked with a substantial amount of original cataloging, (s)he would have a lower production expectation than a cataloger who did primarily copy cataloging.

* Other assignments: When assigning a production expectation, some of the staff in the ASCS had other work assignments, such as computer support, management responsibilities, supervision of student workers, acquisitions work, or collection development responsibilities which had to be considered.

* Professional development and service: When allocating a production expectation to librarian catalogers, the normal contributions of these librarians in the areas of professional development and service had to be taken into consideration.

* Full-time vs. part-time: Not all ASCS catalogers are full-time employees, and the production expectation needed to be pro-rated to reflect the staff member's actual cataloging assignment.

* Non-roman script: This cataloging takes longer when requiring non-roman script parallel fields in the bibliographic record.

* Training: If a cataloger was being trained or was training another cataloger, then this factor also needed to be considered when setting cataloging production expectations.

Each cataloger had a carefully determined production expectation calculated for her/him based on these factors and on consultations with the staff member. It remained important that the cataloger feel that (s)he could meet the monthly assigned production standard.

The quick project planning called for a division of over fifty catalogers. For the most part, the manager of the individual unit determined the expectations, and then met privately with every cataloger in her/his unit. Each unit manager also evaluated the unit's backlogs and drafted a document detailing the following:

* Projections of cataloging progress with current staffing;

* Proposals for increased staffing for the project, if needed, including costs;

* Adjusted cataloging progress projections with increased staffing; and

* Detailed production expectations for each cataloger, explaining how the expectation was determined, including a list of non-cataloging duties.

The nine-month project proved quite successful. It eliminated the backlogs that had developed over the period of the migration and adjustment to a new ILS (Integrated Library System) and when the cataloging of new receipts ceased to focus on retrospective conversion and other projects. It also allowed the catalogers to address any lingering backlogs of difficult-to-catalog materials that had built up over time. Unfortunately, since the time leading up to the project had been so abnormal, with a new ILS and projects taking catalogers away from their normal work, it is impossible to say whether the success of the project resulted from the put-in-place expectations or from management finally allowing catalogers to focus on their normal cataloging responsibilities.

In any case, at the end of the project, a working queue primarily made up of recently published new receipts remained for the catalogers. Occasional exceptions in the form of older gifts and other older acquisitions crept in, but the working queues consisted mostly of recent materials.

The challenge to the managers in the Cataloging Division remained this: how to prevent backlogs from growing again. Catalogers received assurances at the beginning of the project that the production norms developed for the project would no longer be in place at the end. The various units of the Cataloging Division took different approaches for the prevention of future oversized backlogs, and the ASCS decided to develop a set of section-wide cataloging production goals to replace the individually assigned goals set in place for the nine-month Backlog Reduction Project.

Permanent Production Norms for Non-Western Languages

As the ASCS began to meet to discuss the section-wide production standards in the fall of 2002, some effects of the Backlog Reduction Project became apparent. A staff, initially very anxious about production norms, grew very comfortable with the idea of working with this sort of guideline. At the beginning of the Backlog Reduction Project, most catalogers in the ASCS questioned their ability to routinely meet the demanding production expectations for that project. However, when discussing permanent norms for the section, many of the numbers favored by catalogers stood higher than the average numbers in place for the Backlog Reduction Project. One lasting effect of the project focused attention of the cataloging staff on production, and this came out in our discussions on production norms.

For section-wide production norms, staff clearly had to find a different approach than the one used for the Backlog Reduction Project. Cataloger-specific expectations were probably best not "hashed out" in discussions of the whole section. Management determined the expectations for the project. Although the same approach could have been used to devise permanent production norms, beliefs held at the time focused on how management "forced from above" the project production standards. To avoid those ill feelings, it would be better for the new norms to be a set of standards created by the section staff as a whole and for the staff to feel, instead, a sense of "ownership" for the new norms.

The section met several times from October 2002 through March 2003, when it finalized the production norms document for the section. The initial discussions identified the issues that impacted production for catalogers, including:

* The cataloger's level of fluency in the language of the item being cataloged: Some of the Area Studies catalogers specialize in one or two languages, but work with materials in many languages in which they have little expertise. The time spent with a dictionary for these materials lengthens considerably the cataloging time. Also, some non-roman scripts, including Georgian, Armenian, Thai, Old Church Slavonic, and Amharic, can require a great deal of time to transliterate. The staff members who catalog these languages do not really know them well and deal with the languages infrequently, so expertise is difficult to develop. The creation or evaluation of a bibliographic record is very time-consuming when one struggles with the mere transliteration of the script.

* Whether non-roman scripts are input into the cataloging record: At the time of the ASCS discussions, OCLC (Online Computer Library Center, a cooperative online catalog of 53,548 libraries and institutions in 96 countries) accommodated vernacular script input for only CJK and Arabic script records, and these required two separate systems, apart from the OCLC Passport system used by the other catalogers. Only the Arabic language had any sort of automated way of producing a parallel vernacular script field. All of the East Asian languages and Persian required manual creation of the parallel vernacular script fields in those languages.

* Whether the cataloging is done at the PCC (Program for Cooperative Cataloging) level or not: The ASCS staff maintains great knowledge in a wide range of languages. However, not every library has such a staff resulting in a large amount of original cataloging required for many of the languages cataloged in Area Studies. It also means that many of the records available from OCLC are sub-standard and require a lot of work. Because of this, the amount of PCC and/or NACO work done in the ASCS continues to be very high for many of the languages of the section. For example, as of the end of March 2003, the ASCS performed over eighty percent of the PCC work done at the Indiana University Libraries for the 2002-03 fiscal year.

* Whether the cataloger is in training or under revision: At the time of the discussions, ASCS had one cataloger being trained as a PCC cataloger and one cataloger as a non-PCC copy cataloger. This meant that two of the other catalogers served as the trainers/revisers of these two catalogers, so the section had the productivity of four catalogers affected by training. This represented over thirty-three percent of the section.

* Whether the item needs original cataloging or has copy: Original cataloging usually takes longer than copy cataloging, but the impact of original vs. copy can be more complicated. The items more likely to have a record in OCLC continue to be those in the languages which the section handles best: i.e., Chinese, Japanese, Korean, Russian, Polish, Czech, Arabic, Persian, and Hebrew. Naturally, less likely to have copy remain the items in languages which the ASCS catalogers do not know well; as a result, a higher percentage of items requiring original cataloging are even more difficult to process because of the language.

* If the item has copy, whether it has a high or poor quality record: Most of the catalogers who work with copy acquire training at the PCC level, so if they encounter a poor quality record, it receives a lot of attention--NACO authority records and enhancement to the PCC level. Only records already coded as PCC quality go through the cataloging process quickly.

* Other responsibilities of the cataloger: Other assignments in the section at the time included management, collection development, computer support, student supervision, NACO/PCC coordination, and assignments in the Serials Cataloging Unit, as well as the responsibilities of the librarians for professional development and service.

* Whether or not the cataloger is full-time in the section: Staff assignments in the section at the time included 100%FTE, 75%FTE, 50%FTE, and 25%FTE.

With so many different variables influencing each cataloger's meaningful individual productivity, the section saw early on the difficulty in determining a monthly production goal for everyone. Even if original cataloging vs. copy cataloging divided these goals they would be hard to apply across the board. Many of the catalogers do a mix of original and copy, and some do not catalog full-time. Instead of looking at a monthly production goal, the section decided to determine how many items in a specific category a fully trained cataloger should be able to catalog in an average hour.

By choosing the hourly approach, the section eliminated the need to deal with the individual non-cataloging full or part-time differences of the positions in the section. If a cataloger spent a large amount of her/his time on managerial responsibilities that, too, no longer mattered. The production goal applied only to the hours actually spent in cataloging.

Once reaching that decision, the section needed to determine the categories of cataloging for the production goals. The section made its first break-down by separating the cataloging that required the input of non-roman script (CJK and Arabic script only at the time) from that that did not. It believed that vernacular script input added significantly enough to the cataloging time of an item to justify its own category.

The next categories reflected the quality or existence of cataloging copy, broken down in the following manner:

* "Good" DLC/PCC (Library of Congress/Program for Cooperative Cataloging) Copy: This means full-or-core-level DLC/PCC cataloging with an LC-type call number assigned. This type of record can be accepted "'as is" with a few exceptions that require some minimal verification, primarily for local literary call numbers and series treatment.

* Locally Edited Copy: This indicates copy not coded as PCC, but good enough with minor editing that needed to be done only in our local ILS. The national record did not require replacement.

* Non-PCC OCLC Replacement: This involves extensive editing, enough to justify the replacement of the national record, but with the resultant record not coded PCC. There could be a number of reasons for this, among them: the cataloger is not trained for PCC-level cataloging, the cataloger does not feel confident working with the language of the item, or the NACO work required for a PCC-level record is not warranted for the item.

* PCC OCLC Replacement: The poor editing justified the replacement of the national record with the resultant record coded PCC.

* Original (PCC and Non-PCC): Strangely, the section as a whole did not feel that creating original PCC records required a great deal more work than creating non-PCC records. More importantly, however, all original catalogers in the section routinely do NACO work for all of their original cataloging, PCC and non-PCC. Of all of the decisions on categories, this one became the hardest for which to reach agreement. Some of the catalogers who do primarily PCC-level cataloging felt that their original cataloging should have been a separate category from non-PCC original cataloging. In the end, original cataloging of both types received a designation as a single category.

Once the section identified these categories, it finally decided to assign two numeric hourly averages to each of the above categories, one for languages that require non-roman script input and the other for languages that do not. From a management point of view, a very interesting discussion arose. The section could have used this self determination of production norms as an opportunity to seek lower levels for themselves, norms lower than those in effect during the Backlog Reduction Project. Management actually anticipated this result. The project remained challenging, in the sense that staff had to work through a lot of material, much of it undesirable from a cataloger's point of view, in an intense, focused period. Unquestionably, staff felt relief at the end of the project, and sensed the need for a break from the pressure of a production-oriented project. However, a few months later, when we discussed the assignment of numbers to these categories, the ASCS staff assigned actually higher averages than those in effect for the Backlog Reduction Project for most of the categories.

For the category of "Good DLC/PCC Copy," requiring little or no editing or verification, the section decided on 10 per hour for languages not requiring vernacular script input and 5 per hour for those that did. Since the possibility for a DLC/PCC record not already to have the vernacular parallel fields continued, the section believed that these records could need that type of editing, which would be time-consuming.

The "Locally Edited Copy" category had similar goals assigned to it, with the languages not requiring vernacular script input having a goal twice as high as that for the languages requiring vernacular script input: the vernacular script languages had a goal of 3 per hour, while the other languages had a goal of 6 per hour.

As the categories began reflecting more difficult cataloging work, the issue arose of whether an item in a language requiring vernacular input had a lesser impact on the cataloging goal. One reason being that when a cataloger works with sub-standard copy or has to create an original cataloging record, (s)he focuses her/his time and efforts on description and subject analysis. The efforts of inputting vernacular parallel fields are secondary for these records.

Another reason being that the ASCS catalogers in the positions dealing with languages requiring vernacular script input are usually native or near-native speakers of these languages. The catalogers in the languages not requiring the vernacular script input tend not to be native speakers. More likely, they have studied these languages as adults. They are also more frequently asked to catalog a wide-range of languages at a high level, and they often do so with the aid of dictionaries, grammars, and reference works. The task of inputting vernacular scripts is less taxing than the translation tasks associated with cataloging these less well-known languages.

The "Non- PCC OCLC Replacement" category received a 2 per hour quota for languages requiring vernacular script input and a 3 per hour for languages that do not. These assignments begin to show the narrowing of the gap between the vernacular script vs. non-vernacular script distinction. The next category, "PCC OCLC replacement" completes the closing of this gap. The goal is 2 per hour, whether vernacular script is required or not.

In the final category, "Original (PCC and Non-PCC)," the goals reveal a reversal of the order of difficulty between the vernacular script vs. non-vernacular script. Because of the superior native language expertise of the catalogers in the languages requiring vernacular script input, the hourly goal for items in those languages stands at 1.5 for the Arabic script catalogers. Because of the more cumbersome inputting for CJK scripts, the hourly goal for items in those languages remains .5. For original cataloging all parallel fields would need to be input. For the languages not requiring vernacular script input, the hourly goal holds between the two, at 1 per hour, due to the wide-variety of languages, many of which the catalogers do not know well.

The goals, as determined by the staff of the ASCS, are summarized in the table below:

CATALOGING GOALS FOR ASCS

                            Non-Vernacular      Vernacular

Good DLC/PCC                     10                  5
Locally Edited Copy              6                   3
Non-PCC OCLC
Replacement                      3                   2
PCC OCLC Replacement             2                   2
Original (PCC & Non-PCC)         1                   CJK: 0.5
                                                     Non-CJK: 1.5

Comparison of Production under Projects Expectations and Permanent Section Goals

Although staffing turnover makes it difficult to compare production under the goals set for the Backlog Reduction Project with that under the permanent section goals for the section as a whole, some comparisons are possible. For positions that had turnover, one could not tell whether differences in the production of the position were the result of divergences between the project expectations and the permanent section goals or simply dissimilarities between two employees. This study will, therefore, limit its comparisons of cataloging productivity to the nine employees working in their positions for the October-June time frame for both the Backlog Reduction Project and after the implementation of the permanent section goals. Issues such as the same holidays, weather-related absence issues, and routine child-care impact both sets of production statistics for the reduction project and the later permanent norm period, if one compares the same nine months. The section designed the comparison of production statistics for the two periods so that the primary variable continued to be the different production standards in place for the two time periods.

Due to the different statistical accounting kept for the time periods, the data do not match exactly with the established categories. These categories for which production statistics can be determined for the two time periods are:

   PCC Original Cataloging,
   Non-PCC Original Cataloging,
   Replacements of OCLC Record with a PCC Record,
   Replacements of OCLC Record with a Non-PCC Record, and Local
     Editing of Record Only.

The following chart gives the average monthly production for each category per cataloger for the Backlog Reduction Project and for the October-June time frame after the introduction of the permanent section production standards:

Category                               Backlog       Permanent
                                      Reduction       Section
                                       Project        Standard

PCC Original Cataloging                  12.84          27.31

Non-PCC Original                         15.53          23.49
Cataloging

Replacements of OCLC                     27.99          12.88
with a PCC Record

Replacements of OCLC                     18.90          26.00
Record with a Non-PCC
Record

Local Editing of Record                  63.35          86.03
Only (Includes
DLC/PCC)

Total Monthly Average                   138.61         175.01

The above averages do not come close to reflecting the standard for even a single cataloger in the section, but they do provide some interesting information. Despite the fact that the pressure of the Backlog Reduction Project had ended, production stood considerably higher overall with the permanent standards than with the project standards. Only the category of the "Replacements of OCLC with a PCC Record" had a higher production under the project than under the permanent production standards. Production in that category fell by over fifty percent under the permanent section standards. All other categories of cataloging showed a higher production average under the permanent standards than during the Backlog Production Project.

The total monthly average rose nearly forty titles higher per cataloger under the permanent standards than during the project, an increase of over twenty-five percent. Production in the category of "PCC Original Cataloging" more than doubled under the permanent standards, compared to the time of the project, and non-PCC original cataloging production grew by over fifty percent. The average production in both the "Replacements of OCLC Record with a Non-PCC Record" and "Local Editing of Record Only" categories climbed by over thirty-five percent.

Whatever cataloging "shortcuts" the ASCS staff may have implemented during the Backlog Reduction Project, it ended with that project, and most catalogers felt uncomfortable sacrificing quality even for a limited time period. In any case, their cataloging during the nine months under the permanent section standards would have been under the normal rules of quality with no taking of "short cuts." If anything, one should have expected a slight decrease in cataloging productivity when the project finished, and the quality expectations returned to the pre-project standards.

Unfortunately, no pre-project statistics existed to determine whether the project had a significant impact on raising the production of the section. Only statistics compiled after the implementation of SIRSI Unicorn in January 2001 could be used, because the change from NOTIS to SIRSI Unicorn proved so dramatic for catalogers, and the project began late that year. Efforts to learn a new ILS affected the production of catalogers early in the year. Soon after that, management asked them to stop their normal tasks and to work, instead, on retrospective conversion. No significant sample of "normal" SIRSI Unicorn statistics can be obtained prior to the project.

However, a recent random sampling of cataloging statistics from these same catalogers indicates that their productivity remains at a higher monthly average per cataloger now than demonstrated during the Backlog Reduction Project: 150.67 per month now, compared to 138.61 during the project. The responsibilities of several of these catalogers have changed since the implementation of the permanent section production standards, and some of the staff devote less of their time to cataloging. This would explain why today's statistics are lower than those for the nine-month period soon after the implementation of the permanent standards: 150.67 per month today, as compared to 175.01 in October 2002-June 2003. For example, the cataloger of Arabic language monographs now devotes at least twenty-five percent of her time to database management, which was not the case when the ASCS achieved the 175.01 monthly average.

Conclusion

It is difficult to determine why the ASCS staff became even more productive after the intense pressure of the Backlog Reduction Project had ended. Was it because they were able to work on new receipts, rather than on old items that may have been more difficult to catalog? Was it due to the intense stress of the project that came to have a negative effect on productivity at some point? Could it be that the ability of the catalogers to have not only input, but real control over the process of developing their own production standards resulted in standards with which they felt more comfortable? Of course, the likely answer probably lies in a combination of reasons.

The development of production standards for a cataloging unit remains a very complex undertaking, made even more so if the unit works with materials in multiple non-Western languages. The various roles and responsibilities of each individual cataloger must be taken into account, as all non-cataloging assignments will impact productivity. Great care should be taken to involve the catalogers in the development of the production standards. A set of standards "handed down from above" will likely be resented and be seen as unreasonable. However, management should not assume that a set of standards created by catalogers would be less demanding than those established solely by management.

(1) For a recent review of the literature, see Mechael Charbonneau, "Production Benchmarks for Catalogers in Academic Libraries: Are We There Yet?" Library Resources and Technical Services 49: 1 (2005): 40-49.