The Park Library Logo

Newspaper Indexing Policies and Procedures

By Barbara Semonche
Excerpted from News Media Libraries: A Management Handbook , Barbara Semonche, Editor.

Introduction

This chapter offers the reader a historical review of newspaper indexing, a survey of newspaper indexing policies and examples of basic newspaper indexing procedures. While it is impossible to be comprehensive in a single chapter, nevertheless, the intent is to provide more than just an overview. The specific aim is to help those individuals who are interested in undertaking newspaper indexing projects grasp the scale of the task and develop effective strategies for producing a quality index.

This chapter is organized into two sections. The first section deals with the history of newspaper indexing and then turns to general newspaper indexing policies. Also covered are different types of newspaper indexes. The second section focuses on general guidelines for current newspaper indexing operations. Recommendations for retrospective indexing are also provided. A central portion of this section addresses subject-heading control.

History, Policies and Types of Newspaper Indexes

Quick, reliable, comprehensive access to information in newspapers has long been a recognized need. Libraries, historical societies and news organizations have attempted to meet that publicly expressed need in a variety of ways over the years, from bound ledgers, loose-leaf notebooks and card files to keypunch sorters, mainframe computers and personal computers. The major change in the last decade has been the advent of online full-text newspaper databases. Although most of these formats are still in use today, the trend is toward computer-assisted indexing.

With the exception of a relatively few large metropolitan newspaper indexes published and distributed to subscribers, most indexes to local newspapers are not published. They are usually one-of-a-kind projects stored in public or academic library file drawers, notebooks and even shoe boxes. Increasingly they are stored on computer disks. The quality and continuity of these indexes vary considerably. Patron usage and satisfaction with these local newspaper indexes have not been well documented. Factors affecting the success and longevity of newspaper indexing efforts include understanding the fundamentals of newspaper indexing, defining a mission, obtaining qualified staffing, purchasing and maintaining the appropriate equipment, securing sufficient and ongoing financial support and, finally, marketing the index.

Definition of Index

The etymology of the word "index" can be found in several English-language dictionaries. The word is directly derived from the Latin word meaning "that which points out." From Latin dictionaries, the word has more complex meanings, for example, "one who informs or discloses" or the "forefinger" or "of things, that which informs, a sign, token" and even "a touchstone." For a brilliant discourse on the etymology of the word "index," consult Hans H. Wellisch's book Indexing from A to Z.1 He also offers some excellent advice for neophytes considering the indexing of newspapers. For our purposes we will accept that an index is a systematic guide to items contained in or concepts derived from a collection. These items or concepts are represented by entries arranged in a known or stated searchable order, such as alphabetical (author and subject or topic and key word), chronological or numerical.

A newspaper index should not be confused with a book index. Book indexes usually index the content of individual monograph publications. Typically they appear at the back of the book and are alphabetical, including subject and name terms. Books are separate, discrete entities with a unified approach in content and style. Newspapers are ongoing publications with episodic coverage of people and events written in varying styles. The most significant difference is the frequent and ephemeral nature of newspapers. That newspapers are short-lived, fast-paced, linguistically fluid and overfull with small and seemingly insignificant information is testimony to their function as a mass medium. Philip Graham, former publisher of The Washington Post, referred to newspapers as the first rough draft of history.2 Obviously, the final edition is not yet written. Harry Friedman's book Newspaper Indexing offers this description of a newspaper index:

A newspaper index, because of its extreme usefulness, must be made to fit the time, the place and the occasion. It must be made to suit the needs of its creators and it must meet the demands of every type of user. . . . To be accurate and comprehensive, and, more important, to be objective in a scientific sense, the newspaper index should not interpret news and opinion according to the beliefs of its creators-it must reflect only what actually exists.3

History of Newspaper Indexing

A good deal of indexing of books and periodicals preceded newspaper indexing. One book, How to Make an Index, written in 1902 by Henry B. Wheatley, is highly recommended. It is a unique and admirable guide to the fundamental process of indexing. It is also witty as well as informative. By way of examples, consider the following chapter headings: "The So-called Evils of Index Learning," "Amusing and Satirical Indexes" and "The Bad Indexer," followed, naturally, by "The Good Indexer." The book ends with discussions on the practical elements of indexing, classifying, alphabetizing and the like.4

The dean of modern periodical indexing is William Frederick Poole, who in 1848 published an Index to Subjects Treated in the Reviews and Other Periodicals. In 1853 he improved his first effort and retitled it Index to Periodical Literature. Poole's Index is considered by some as the forerunner of the printed The New York Times index. Undoubtedly, it gave considerable impetus to the indexing movement in the United States. However, European newspapers, particularly British ones, were in the forefront of early newspaper indexing.

Newspaper indexing, while not an ancient art, is hardly new. Still, there remains some difficulty in differentiating this process from book or periodical indexing. Wheatley found that a 1601 translation of Pliny's Natural Historie of the World had an index. Speed's History of Great Britaine, published in 1611, also had an index. However, these indexes were more closely aligned to abstracts than typical "pointer" indexes.5 The Lathrop Report on Newspaper Indexes lists the Maryland Gazette index (1727-46) as the first of its kind in the United States.6 The Maryland Gazette index, consisting of 377 typewritten pages, is bound in four volumes. Lathrop also identifies Canada's earliest newspaper index in 1787 as that of the Royal American Gazette & Intelligencer of St. John and Prince Edward Island. What is noteworthy about these indexes is that they have survived. In this current era of high technology, when the equipment required to retrieve or display retrospective information is occasionally malfunctioning and costly to repair, it is satisfying to be able to refer to the simple written record.

Sometime in the 1850s, newspaper indexing actually began to expand and develop along the lines of modern indexing. For more than 100 years, attempts had been made to index the press of the period. The Spectator and Tattler, London's literary gossip and sa tirical newspapers, were indexed as early as the turn of the eighteenth century,7 and, of course, the London Times has been indexed as far back as 1790. Robert Desmond's book, Newspaper Reference Methods, contains a chapter with useful information about e arly newspaper indexes in the United States.8 Since then, other newspapers have been indexed for varying periods of time. These indexes also vary in scope and format. Some are handwritten notebooks or ledgers with just a listing of obituaries. Others are on typed catalog cards and include articles on local events as well as personalities. Some include annotations. A few index national and international events. Photos, however, were rarely indexed.

The most prominent and perhaps most widely used newspaper index in the United States is The New York Times Index. Established in 1851 as a handwritten guide to the contents of the newspaper, it continued in this form (with the exception of a missing index for 1859) until 1860, when the first printed index, only two pages, was issued. A three-year interval elapsed before the next issue of the index, but it continued without interruption until 1905, when the printed index was suspended.9 In 1913 the present The New York Times Index started and continues to the present date. Online periodical abstracts and selected full-text articles were offered to the public via The New York Times Information Bank in the mid-1970s. In 1983 marketing and distribution rights to this online service were sold to Mead Data Corporation, which gave birth to NEXIS, a full-text online news database companion to Mead's law database, LEXIS.10

Almost simultaneously with the publishing of The New York Times Index, the New York Herald began its own index. Robert Bligh, one of America's first newspaper librarians, started indexing all the way back to the first issue in 1835. Consisting of entries in longhand on large ledger sheets, Bligh's undertaking was considered one of the most extraordinary tasks in the newspaper library field. From 1880 until 1920, when the Herald and the New York Tribune were merged, the index was under the direction of sev eral librarians. The index, which consists of 1,742 volumes, is written entirely in longhand with the exception of a period from 1908 to 1920 when it was typewritten and four carbon copies made.11

National, Regional and State Newspaper Indexes

As early as 1893, the American Library Association expressed interest in and the need for newspaper indexes, particularly a national newspaper index. However, R. R. Bowker, an advocate of newspaper indexes, saw substantial difficulties in developing a na tional newspaper index. He wrote: "To index newspapers of the past as Dr. [William Frederick] Poole has indexed periodicals . . . would be practically an impossibility, valuable beyond measure as such an index would be."12 Nevertheless, the Annual Literar y Index followed the next year under Bowker and William I. Fletcher. The plan was to index eight to ten newspaper dailies covering news of general interest that would be indexed by dates and not as references to any one paper. The index was to be planned as a cooperative venture, the actual work being done by librarians in the designated cities. The next step was to forward these indexes to a central site where a composite index would be created. The Annual Literary Index was issued at regular intervals u ntil 1904.

The concept of a news index that would serve as a key to all the newspapers of Michigan was conceived in 1920 by Stuart H. Perry and J. S. Gray of the Adrian Daily Telegram. The concept took form in 1924 at the fall meeting of the Michigan Press Associati on. The Detroit News, the Detroit Free Press and the Grand Rapids Press were selected as the papers from which a statewide index would be made. The Adrian Telegram was to supervise the actual indexing and publish the periodical issues of the index. The in dex was started in February 1925. It was suspended in late 1926. For more than a year the index was furnished to 25 subscribers in two forms, temporary sheets in newspaper form every week and bound book form at the end of every quarter. The subscription p rice, between $100 and $150, became too expensive, and subscribers dropped out.13

This idea of a composite newspaper index has not died. It returns from time to time as public and academic librarians struggle to unify, standardize and control subject access to a region's or a state's newspapers. The efforts, well meaning as they are, a re doomed to fail over time. There is the mistaken concept by many not directly affiliated with the newspaper industry that all newspapers are somehow alike. Further, it is assumed that it would be more convenient to use the identical subject headings and the same article selection policy to index a dozen or more papers. Not only is such an approach naive, it turns out to be impractical and unrealistic. Even if by some chance such a project could be launched, the resulting index would be too unwieldy to g ather into a composite volume, particularly for the annual cumulation. On the other hand, a logical attempt by the indexing committee to control the index size by severely restricting article selection and relying upon only the most general subject terms would defeat the major purpose of an index. Library patrons would be disappointed and frustrated with such ineffective access to material published in their local paper.

A more difficult-to-anticipate indexing problem, especially for composite projects, is the decision-making process for adding new subject terms, deleting old ones, dividing subject terms and selecting cross-references. The communications problems emerging from these daily occurrences would boggle the mind and overwhelm the spirit of even the most dedicated indexers. The alternative of having one central indexing operation undertake the task of indexing several newspapers simultaneously reduces the procedu ral problems somewhat. Nevertheless, costs, quality and timeliness remain difficult to manage on multiple-newspaper indexing projects, even given the advantages of high-powered computers and effective indexing software.

A far better approach for librarians interested in launching a newspaper indexing project would be to treat local newspapers individually and index appropriately for the users in each paper's circulation area. Scholars and other patrons eager to search a topic through several of a state's papers may have to adjust to different terminology. That problem can be mitigated with abundant cross references, useful scope notes and a clearly expressed article selection policy. Foremost in the mind of indexers shou ld be the eternal questions, "What types of articles would the public seek?" "How would someone, anyone ask for this information?" "What terms would they use?" To avoid these questions is to discover that the public will not use what is not useful.

Early Newspaper Card Indexes

Prior to card indexes, which could be easily updated with new entries, libraries relied upon loose-leaf ledgers with handwritten or, occasionally, typed entries. The problems with expanding subject and name headings while maintaining chronological order were manifest. A better system had to be found.

The origin of newspaper card indexes is unclear. One of the chief claimants is the San Francisco Chronicle, which started its index in 1864 and continued until the disastrous fire of 1906 destroyed all its records and files. Partial substantiation of the Chronicle claim is contained in a newspaper report of the dedication of the newspaper's building in 1879. A reference was made to seeing "in the three hundred tin pigeon holes along the blank wall of a narrow room what was doubtless the germ of the index card system."14

In 1927 an 86-item questionnaire about newspaper libraries was prepared by the Newspaper Group of the Special Libraries Association. The survey was mailed to 380 newspapers. Seventy replies were received. Among the questions asked was one about indexing. Twenty-nine librarians reported indexing their newspapers, 15 on cards and 4 on long sheets. The remaining ones were apparently too closely allied with the clipping files to be called true indexes.15

With the advent of the Works Progress Administration in 1935, newspaper indexing enjoyed rejuvenated interest. The WPA was called upon to provide funds for the digesting and indexing of newspapers. More than 20 newspapers from across the United States lau nched projects covering editions from 1728 to 1940. In addition, some collections of newspapers owned by historical societies and schools were indexed and microfilmed.16 Friedman's book on newspaper indexing lists some nineteenth century newspaper indexes .17 Most are incomplete and many have long been discontinued, but some can still be consulted by checking newspaper archives in large public and research libraries.

The Lathrops, Norman and Mary Lou, created the Newspaper Indexing Center in Flint, Michigan, in 1969.18 They wanted to create the first comprehensive, descriptive bibliography of all published and unpublished newspaper indexes in the United States and Can ada. The couple's efforts resulted in The Lathrop Report on Newspaper Indexes, first published in 1979.

A detailed listing of valuable newspaper archives, including indexes, can be also be found in the 1991 publication Untapped Sources: America's Newspaper Archives and Histories prepared by the Gannett Foundation Media Center for the American Society of New spaper Editors. The interested reader should also consult several excellent journalism bibliographies, notably Jo Cates' 1990 Journalism: A Guide to the Reference Literature and Eleanor Blum's and Frances Wilhoit's 1990 Mass Media Bibliography: An Annotat ed Guide to Books and Journals for Research and Reference.

Newspaper Index Types

Barbara Newcombe, in a May/June 1991 article in the American Society of Indexers' Newsletter, refers to the differences between producing an index designed for the use of news staffs and one designed for use by the general public. She further comments up on the differences between indexing for a computerized bibliographic database and a full-text database with free-text retrieval.19 In the best of all possible worlds, an excellent index (accurate, current and comprehensive) should be able to meet the need s of the most penetrating investigative reporter and the most casual public citizen browser. It is a mistake, however, to classify information needs of the public in general terms as if that group required only "generally indexed" information access. Publ ic reference librarians can offer eloquent testimony to the almost infinite variety, depth and urgency of querying clients, from students, scholars and local historians to genealogists, freelance writers and private investigators.

Beyond the simple fact-finding mission of clients' queries may lie a less obvious, more elusive need. Newcombe recognizes a reporter's special need for retrieving stories on different levels. News staff, on occasion, need to locate stories according to su ch terms as "color pieces," "reaction stories," "analysis," "wrap-up" and "backgrounders," as well as a score of other difficult-to-define specialty terms.20 However, it should not be that difficult or more time-consuming or costly to enhance stories with these terms and others (for example, feature, profile, series, chart, rankings, and so on), as long as indexers know how to properly identify such articles. Certainly the addition of these enhancement terms does not eliminate the need for full indexing u sing additional appropriate terms to aid retrieval for a variety of inquiries. These specialty terms should in no way detract from the public's access. Further, on some occasions the information needs of both news staff and the general public may overlap. As a rule, unique search terms enhance precision retrieval even in precoordinate indexing typical of printed formats.

At one time, early in the development of full-text news databases, it was thought that additional enhancing or indexing was unnecessary. After all, one could search not only on the headline, dateline and byline, but the entire article. Surely that was bet ter than an abstract or simple bibliographic entry. In short, what more could be needed? The passage of time, expanding news coverage, investigative reporting and the vagaries of news reporting styles have revealed how wrong that assumption was. Some addi tional intellectual effort and news article analysis are necessary to link appropriately all the myriad, fleeting and unstated associations among a never-ending stream of stories about people, events and institutions. News librarians were discovering that as their databases grew and reporters came and went, a high level of information retrieval precision was difficult to maintain. The quality control of online full-text databases is definitely improving with enhancement of effective indexing techniques, a dditional costs notwithstanding.21

Local Newspaper Indexes

Newspaper Library Indexing Projects

Typically, newspaper libraries do not index their papers for their own staffs. The few that do are exceptions. Traditionally, reporters, editors and photographers have relied upon their libraries' clipping files and photo collections rather than publishe d indexes. These files are arranged by personal name and subject and usually contain all that is needed.

Indexes are time-consuming for news staff to use, requiring several steps before the information is captured. Immediate, full-text access (clip files or online database) is preferred. However, not all newspapers have in-house libraries, or, if they do, th ey are so lightly staffed that maintaining a labor-intensive, space-devouring clipping file has too many disadvantages. Local news indexing in these circumstances is a reasonable option.

One example of a newspaper library generated-index is described by Colleen Daze, librarian with the Schenectady Gazette.22 She relates her experience facing an expanding clipping collection and no space for growth. Daze's indexing procedure involved typin g prepared index entries into her newspaper's front-end electronic typesetting system. A reporters' byline index was also prepared. Copies of the index are made available to the paper's in-house news staff plus reporters at ten news bureaus. The general p ublic gains access through an index copy sent gratis to the public library. The current index is online, but previous years' indexes are available in paper format.

Among smaller and midsized newspapers, the number of newspaper libraries preparing, publishing, marketing and distributing their own indexes to public subscribers is small indeed. Notable among these few is a relatively long-term index to three small (com bined circulation 65,000) Southern newspapers. It is IND-EX, published by the Durham Herald-Sun, Inc. From April 1981 to February 1990, the Herald-Sun library with a staff of two to three indexers prepared monthly newspaper indexes and annual cumulations for 15 paid subscribers. This indexing project was initiated following research undertaken by a Herald-Sun library intern, Gloria Colvin.23 When the project was launched in-house, additional library staff were employed to maintain the index. It was a sepa rate project from the in-house clipping file operation. The news staff did make regular and frequent use of IND-EX, particularly for byline reports, but, as expected, it was the clipping files, complete and in date order, that the news staff turned to for faster response time. After nine years the Herald-Sun Index was reluctantly discontinued as a result of company-wide budget cuts. A procedure manual of this new indexing operation is available from this chapter's author.

Public and Academic Library Indexing Projects

Public and academic libraries have a long history of starting to index local newspapers (on cards or in ledgers) only to become overwhelmed by the enormity of the task. The investment of money, the need for managerial support, the twin problems of staff shortage and training, the questions of whether to market the index or not, the difficulty of maintaining consistency and currency, and the tribulations of selecting and keeping up-to-date with equipment modifications and software enhancements combine to frustrate even the most dedicated public librarians. Hence news indexes may be started but are quickly found to be costly, time consuming and difficult to maintain. As a result, they are likely to be just as quickly discontinued. However, even such incomp lete indexes continue to be valuable.

Many, but not all, state libraries have printed guides to their newspaper indexes. Public and academic libraries in the states make every effort to keep as current as possible with their newspaper indexes. For an overall, albeit somewhat dated, view of ex isting newspaper indexes, consult the works of the Lathrops,24 K. D. Sell,25 and Anita Cheek Milner.26 A more recent description of a college library computerized newspaper indexing project was delivered in 1992 paper by Candace Peterson.27

Occasionally groups of libraries have banded together to undertake news indexing projects in anticipation of sharing responsibilities and costs. Interestingly, this approach seemed to complicate matters rather than simplify them. Questions about newspaper indexing policies and procedures proved difficult to identify and resolve over time and distance. The policy and procedural considerations start with determining the best computers, peripherals and software available and then spiral into a myriad of othe r decisions. How long can the project continue without maintenance and upgrades? Will the equipment invested in be superseded by newer versions? What is the cost of conversion to newer computers? Bear in mind that the handwritten or printed news indexes o f 250 years ago are still usable. Can we forecast that the indexes, thesauri and authority files stored on the computers using the software purchased on today's market will be available even in ten years time? How will the vocabulary be controlled? What a bout updates and cumulations? What will be the marketing approach? Who will train and supervise data-entry staff? The questions seem endless, the answers elusive. Following the lead of others helps. Two examples of cooperative indexing are described in ar ticles by Carol Cole28 and by David Nelson.29

Newspaper Indexing Software

Until 1982 there were no indexing programs for microcomputers. In the past decade there has been a marked increase in the number and quality of indexing software programs. For the most part, however, they have been targeted for the "back-of-the-book" ind exer rather than for newspaper or periodical indexers. A few of these innovative programs can meet some newspaper or periodical indexing requirements. Experimentation with demonstration disks offered from these indexing vendors will help in selecting soft ware.

Linda Fetters prepared A Guide to Indexing Software for the American Society of Indexers.30 It is a valuable review of existing microcomputer indexing software. While most of the software programs reviewed were more appropriate for "back-of-the-book" inde xing, this guide is still the best place to begin an investigation into the market. Essentially, there are two kinds of indexing programs available. The first is a text-embedded style designed to work with word-processing files. The second type works like electronic index cards. It is possible, with some accommodation, to create newspaper indexes with existing software. Currently, there are no software programs designed exclusively for newspaper indexing available at a moderate price. The full-text online systems offer some indexing capability, but they are usually attached to a newspaper's front-end electronic typesetting system. Fetters's guide provides authoritative descriptions with illustrations of such software programs as Cindex, IndexAid2, Indexer 's Assistant, Indexit, INDEXX, Macrex Plus, wINDEX and Micro Indexing System. She also includes a publications list from the American Society of Indexers covering indexing bibliography, proceedings and where to find advice for selecting a computer for ind exing. Many of these programs have demonstration disks available for interested clients at nominal cost.

Hypertext is another type of indexing. It was first developed on mainframes and minicomputers in the 1960s and 1970s. It was not until about 1987 that it appeared on microcomputers. Examples include Apple Computer's HyperCard, IBM's Textbank/PC, Houdini f rom MaxThink, Software Garden's KnowledgePro and Eastgate's Hypergate. There is debate about what are so-called true hypertext programs. A full discussion of this topic is beyond the scope of this chapter. Information is readily available in local librari es.

On the cutting edge of experimental indexing software you will find such topics as automatic indexing, artificial intelligence and expert systems. The interested reader is invited to explore these topics in other sources than this chapter. At this time we can report that expert systems are intended to capture human expertise and encapsulate it within a computer system. The American Petroleum Institute in the 1980s designed an expert-like system for automatically indexing abstracts for the API database. A thesaurus developed over 20 years of indexing was used. The goal was not only to reduce the cost of indexing but to increase the efficiency of searching the database.

The library literature offers several creative examples using word-processing programs and database-management software for indexing newspapers. Ted Hunsberger described using dBase III+ to index the Fort Wayne Journal-Gazette.31 Michael Cramer and Mary M arkland used the Pro-Cite software package to index several papers, including the Roanoke Times and World News.32 Charles Anderson listed the advantages of using Savvy PC as a database for several reference purposes, including indexing the Northbrook Star .33 Michael Knee briefly outlined his experience using three bundled software packages, WordStar, Personal Pearl and SuperCalc, to index the Albany Times Union.34 Celia Wall described using an IBM mainframe computer and a text-formatting program, SCRIPT, to index two local newspapers in Kentucky.35 Jean Koch offered a description of several database software packages appropriate for indexing on microcomputers.36 Note that these programs may have been modified or superseded. Fetters's guide to indexing sof tware is more recent and comprehensive.

The advent of personal computers and relatively easy-to-use indexing software packages has made newspaper indexing more attractive, but not without costs and certainly not without the need for continued top-management support. Indexing local newspapers is not something librarians can do in their spare time. Relying upon volunteers, eager and inexpensive as they are, presents special problems in training and scheduling. Consequently, widespread, convenient, accurate, timely and comprehensive access to loca l news published in local newspapers is rare. However, newer, simpler computer-assisted indexing of newspapers is making the long-term objective possible and even economical.

Published Newspaper Indexes

While corporate and research libraries make extensive use of costly online full-text news databases, other major libraries as a rule do not. How to pass on news database search fees to students, faculty and scholars who are accustomed to no-fee access is a severe problem for public libraries and even large academic libraries. As a consequence, there is still heavy reliance by these libraries upon the relatively low-cost published newspaper indexes. Typically The New York Times Index, The Washington Post Index, The Christian Science Monitor Index, the Atlanta Journal-Constitution Index, The Wall Street Journal Index, USA Today Index, as well as international newspaper indexes are standard news reference tools. Descriptions of some of these indexing projec ts can be found in works by John Rothman37, Alan Greengrass38 and Jessica Milstead.39

Not to be overlooked among published newspaper indexes is the Index to Black Newspapers, which started in 1977. It is compiled quarterly from the final editions of at least eight prominent black newspapers. The current list includes Afro-American (nationa l edition), American Muslim Journal (Chicago), Amsterdam News (New York), Call & Post (Cleveland), Chicago Defender, Daily World (Atlanta), Michigan Chronicle (Detroit), and Sentinel (Los Angeles). The index covers local, regional, state, national and int ernational news with subject and personal-name access.40

Also in the area of special newspaper indexes is the Alternative Press Index, started in 1969, which covers purportedly alternative and so-called radical publications.

The Canadian News Index is another published newspaper index. Started in 1977, it continues with monthly indexes and annual cumulations. Coverage extends to the Calgary Herald, Toronto Globe and Mail, Halifax Chronicle, Montreal Gazette, Toronto Star, Van couver Sun and the Winnipeg Free Press. The subject index covers governmental and corporation names, while individuals appear in the personal-name section. References are arranged chronologically. Online searching for articles appearing in several of thes e newspapers since 1981 is available on several database services.

In 1992 the United Kingdom daily paper, The Independent, announced a new monthly official index. The Independent Index for the first six month of 1992 will be available in two paper-bound quarterly cumulative issues from July. Subsequently, the monthly is sues will be cumulated annually into a cloth-bound volume. "The index is prepared by a team of indexers of Research Publications International based in Reading, UK. The aim is for a three-week turnaround from keying to customer, with access to news, edito rials, features, financial information, captions, letters, and obituaries."41 Research Publications International also publishes indexes to The Times, Financial Times and Le Monde.

News digests such as Editorials on File (from 1970) and Facts on File (from 1941) provide useful access to national and international news. Editorials on File reprints leading editorials from over 150 newspapers. It is indexed by subject. Facts on File of fers factual summaries of national and world news. The advantage of this weekly index is that it provides access to the contents of newspapers before their own indexes are published. It is also available online.

CD-ROM products such as Newspaper Abstracts and the full-text ProQuest Newspapers Ondisc are alternative methods of gaining subject access to newspapers. They are not cheap. While the hardware costs are declining, the updates on discs are not. With a larg e clientele these products can be useful and cost-effective.

Information Access produces InfoTrac with automated bibliographic recovery. InfoTrac contains indexing to nearly 800 periodicals, including a year of The Wall Street Journal and two months of The New York Times. The information is on optical disc and is r etrieved by patrons using PCs. While InfoTrac fills in the lag-time gap from printed indexes, it does have some retrieval problems. One current drawback is that search terms cannot be linked by Boolean operators. For large libraries, the big investment co sts of CD-ROM products may be manageable.

NewsBank offers libraries a package of newspaper articles on microfiche with indexing. Articles are selected and microfilmed from selected newspapers published in all 50 states. Access is through rather broad index subject headings. Almost 100,000 items a re added annually, but no one paper is completely indexed. NewsBank indexes are available on an automated, self-contained system, NewsBank Electronic Index.

The availability of these indexes in varied formats, while useful, does not diminish the urgency for subject access to local newspapers. Printed indexes to local news continue as a vital information need.

Online Full-Text News Databases

Newspaper Databases

In recent years, online full-text news databases of nearly 100 large metropolitan newspapers, several dozen smaller newspapers and numerous national and international wire services have been made available at a price. While this number will continue t o grow, the reader should bear in mind that there are over 1,600 daily newspapers and over 9,000 nondaily papers in the United States. Even as the number of online newspaper databases continues to grow, it is small in comparison to the total number of new spapers published in this country. It is doubtful that all, or even the majority of, papers will be in electronic retrieval format accessible by the public in the near future. Also, full-text databases do not include everything published in the paper. Thi s is yet one more reason for realizing that the need for local newspaper indexing does not diminish with these online full-text news databases. If anything, the need grows.

Some online vendors offer gateways to news databases. H & R Block's CompuServe, for example, provides access to 850 databases, 48 newspapers including AP wires and a specialized "clipping service." The principal database vendors are Mead Data Central with NEXIS/LEXIS, Knight-Ridder's VU/TEXT and DIALOG, DataTimes and Dow Jones News/Retrieval. DIALOG offers full-text access to over 80 newspapers and magazines. We can expect to see a merging of DIALOG's newspaper files with VU/TEXT as a result of VU/TEXT's purchase of DIALOG in 1989. VU/TEXT currently offers over 70 newspapers full-text. NEXIS/LEXIS provides more than 16 wire services, dozens of magazines and newsletters, plus over 100 newspapers and business journals. In short, these database vendors suppl y an extraordinary amount of information, most of it since mid- to late 1970s. There are other online services. Check with Gale Research's Computer-Readable Databases: A Directory and Data Sourcebook for an up-to-date list.

For discussions about the early status of online news retrieval systems, consult works by Stephen Smith,42 Carol Tenopir,43 and Reva Basch.44 A very useful and recent article about full-text, online newspaper databases was written in 1991 by Dorothy Ingeb retsen and Thomas Lutgen. The authors compared and contrasted varying strategies for searching the Los Angeles Times on DIALOG, VU/TEXT and NEXIS.45 Anne Mintz46 and Reva Basch47 offer valuable insight into searching and quality control of full-text datab ases.

The authoritative source for online, full-text newspaper databases is a guide prepared by the News Division of the Special Libraries Association. It is designed to guide news librarians, their managers and data-processing staff in the development of in-ho use electronic clipping files.48 For an update consult Judy Metcalf's chapter in this book on electronic news libraries.

Broadcast News Databases

Radio and television stations are awakening to the value of convenient access to information previously published and/or broadcast. For in-house indexing operations, check "Television News Libraries" by Jeannette Kopak later in this book.

Printed indexes to broadcast news are few, but they are available in larger academic and research libraries. The Television News Index and Abstracts: A Guide to the Videotape Collection of the Network Evening News Program in the Vanderbilt Television News Archives reaches back to 1968. A monthly publication with quarterly cumulations, it has author, title and subject indexes for the combined coverage of ABC, CBS and NBC. There was no weekend news recorded until May 1970. Local preempting of the evening ne ws for other programs, which occasionally occurs, causes some unavoidable gaps. The Macneil/Lehrer Report: Broadcast Review and Index has been available since 1979.

Recently, commercial databases have begun to include the full text of news and public-affairs programs from television and radio. One of the vendors, Burrelle's Broadcast Database, offers timely, comprehensive access to ABC, CBS, NBC and FNN as well as Na tional Public Radio. Network shows are not included. Unfortunately, there is not yet indexing for personal or company names. Searching is further complicated by the current limitations of the User Guide and its lack of a thesaurus. If Burrelle's Broadcast Database does well in the marketplace, doubtless these problems will be properly addressed. Other broadcast databases include NEXIS's full-text transcripts of Public Television's "The Macneil/Lehrer Newshour" and DELPHI's full-text transcripts of "Nova" and "Frontline."

Online searchers have several sources for keeping current with the increasing number of electronic newspaper abstracts and full-text news databases. Access to such journals as Database Searcher, CD-ROM Librarians and Computers in Libraries is the mark of the true information professional.

While a number of indexing case studies have been reported in the literature in recent years, including newspaper indexing projects, there has been little in the way of research on the topic. Most of what has been done focuses upon postcoordinated indexin g or interactive searching done online. Two helpful articles, one by Tefko Saracevic49 and the other by Louis Gomez, Carol Lochbaum and Thomas Landauer,50 include valuable bibliographies.

GENERAL GUIDELINES FOR NEWSPAPER INDEXING

Indexing Goals

Whether a newspaper index is established to complement a news library's in-house clipping files or to be published and distributed through subscriptions to a general clientele, the principles of well-constructed indexing should prevail.51 If both types o f clients are to be served, interviews and surveys with prospective users are definitely recommended. Such interaction is vital to determine the indexing policies, scope, content and arrangement before the project is launched. Careful planning allows the demands of both groups to be accommodated. Do not assume, however, that an index designed to meet the needs of news staff will automatically satisfy public subscribers. Certainly the reverse will not hold.

Most frequently overlooked by eager newspaper indexers is research to uncover long-out-of-date, but still-existing local news indexes. Even more surprising would be the discovery that someone else is indexing the same paper simultaneously. A proper survey should reveal such information.

Generally speaking, the essential functions of an index are to identify and locate specific items in a collection. An index may also indicate the nature and scope of the collection, furnish additional data (annotation) and correlate the language of the us er with that of the index (through a thesaurus or subject-authority list), as well as indicate the relationships between items.

The quality of an index is determined largely by the accuracy and completeness of individual entries and the consequent adaptability of the index to the shifting needs and changing language of its users over periods of time. The success ratio of the searc h strategies applied to the index is a critical factor. Quality is also affected by such considerations as cost and time limits of the operation, size of the index and skill of the indexers. Indexes will not satisfy all criteria equally. User requirements must be balanced against production capabilities.

Initiating an Index

The success of an index is largely due to thoughtful and thorough planning. Intelligent decisions about the most suitable type of index, its scope and form, and the method used to produce it can be made only after specific questions have been satisfactor ily answered. Consider the following:

  • Who will use the index?
  • What types of information will users seek?
  • How can information be displayed for greatest usability?
  • Will the index function with clipping files, bound volumes, microfilm or by itself?
  • How long will daily indexing take?
  • How much information will be indexed?
  • What types of articles will be selected for indexing?
  • What types of articles will not be selected for indexing?
  • How much computer storage space is needed?
  • How will subject-heading control be maintained?
  • How can the information be coded and stored efficiently?
  • How will the editing process (for index accuracy and consistency) be managed?
  • How will the files be organized and stored?
  • Will the index be retrospective or current or both?
  • What might be the future uses for the indexed information?
  • How frequently will the index be printed and distributed?
  • What capital investment will be required for the indexing project?
  • What are the anticipated annual operating costs?
  • Has a marketing study been undertaken to determine the potential for index revenue from paying subscribers?

The answers to a few of these questions may be readily apparent. Some may seem less obvious. Most will require study. A few basic procedures are useful in investigating such questions, making wise decisions and avoiding errors. Learning from the experienc es of others is valuable in reducing errors and saving time. Examining a variety of types of indexes can help in determining the format and type of index that best meets the needs of a particular newspaper. If an index is to be of maximum value, the needs of its users must influence the selection of items to be included. A user survey can determine who will use the index, what type of information should be included in it and what its scope should be. If the index is to be marketed, this survey can determi ne the market area and identify potential subscribers.

Establish guidelines and policies to clarify what types of items will be included, the depth of indexing and the way in which different types of items will be indexed. Start a preliminary subject-authority list to test subject-heading consistency and cont rol. Index editing requires careful attention in order to maintain quality control. All these procedures can be generally outlined to provide an idea of work flow and to determine staff requirements. Later, they can be expanded into a procedures manual.

Regardless of how the information will be stored, it is important to know how much information will be included and what its approximate rate of growth will be. Records of the number of articles indexed per day and the number of indexing terms used per ar ticle kept over a period of several weeks can be used to compute an average number of articles and indexing terms. Projections can then be made about the size of the files in a week, a month or a year, and estimates can be made of the space needed to stor e the information.

Maintenance of an index costs money. The investment is balanced to some extent by the value of its contribution and, of course, by its revenue-generating potential. Nevertheless, to maintain a useful index requires money in addition to talent. Before star ting an index, a commitment to continued financial support should be established. Preliminary budget estimates can be prepared once the indexing approach and method have been decided. These estimates should include equipment, supplies, printing, promotion al costs and staff salaries. Some expenses will be one-time capital expenditures, while others will be ongoing. Comparison shopping for equipment and maintenance services is recommended.

Indexing Approaches and Methods

Before embarking on an indexing project, the indexer must determine the indexing to be done, begin a subject-authority file, decide on the indexing approach and determine the level of specificity desired. Guidelines should then be written to assist index ers in maintaining consistency from article to article over time. These guidelines should identify the circumstances under which personal names, and names of organizations, places, events, things and other items should be indexed for particular and relate d articles.

One consideration in establishing an index is the depth or thoroughness with which an article should be indexed. The possibilities range from listing each article under just a single heading to indexing several selected key words or to making an article a vailable under almost every word found in its context. In all approaches, the goal is to make the article readily accessible within limits imposed by space, time, personnel, equipment and money.

Small indexing projects (most often those that record citations in card files) limit the number of subject headings to one or two general terms. In such situations, the indexing is done manually, and time prohibits assignment of several specific subject t erms and personal names to each article. Even so, as these relatively small files grow, they will reach a point at which their volume will make searching difficult.

Other indexers manually assign multiple subject headings to articles based on important names or words that either appear in the article or are implied by it. The number of terms used is dependent upon the contents of the article: three or four relatively specific terms per article are not uncommon, while some local articles require even more. These terms are in addition to the personal names assigned to the articles.

Retrospective Indexing

Indexing newspapers decades old, from crumbling bound volumes or from faded microfilm, is a special challenge. A pilot project should be undertaken to understand the unique and sometimes puzzling elements involved. Prepare a preliminary index record form to use with each article indexed. Select random weeks at intervals of five or so years. Note reporting style and newspaper format. Are bylines and datelines included? Are wire services identified? Note variant spellings of persons and institutions. It wa s not unusual for first names to be omitted in news copy, so name verification will be troublesome. Attempt to use the same subject headings from the current newspaper indexing project, if any. Be prepared to enter new terms even if they have changed in l ater years. Rely upon cross-references and scope notes to link related terms and topics.

Recognize that older papers have faded over time. Typeface was much smaller years ago and tends to impede legibility now. Parts of pages or whole sections may be missing. Even more confusing may be a lack of follow-up articles on what seem to be major sto ries. Finally, newspaper design and format vary from year to year in older papers, making it difficult to predictably locate certain types of stories. If retrospective indexing is difficult, the editing process will be even more challenging. This is a ver y slow, painstaking process. Be certain that the result will be worth the investment of time and money before actually launching such a project.

For a case study of both retrospective and current newspaper indexing, consult an article by Bobbie Vierra and Tom Trice.52 This project encompasses two sets of newspapers-selected papers from 1857 to 1900 and a current daily paper. The two indexes have d ifferent scopes and are housed separately but share a common subject-authority file and format. An indexer's manual was developed to provide a portable authority file and to instruct indexers on indexing principles, scope rules and difficult components. T he Napa City-County Library staff and a few volunteers do the indexing.

Thomas Armstrong and Janice Fennell offer another case study in retrospective newspaper indexing. This project, initiated at Georgia College, developed an index and abstract for two Georgia papers from 1820 to 1982.53

Subject-Authority File

There is some difference of opinion about the appropriate name for this concept. The terms suggested include subject headings or subject-authority list, controlled vocabulary, thesaurus, topics, key words, terms (main, broader, narrower or related terms) , descriptors and/or identifiers and enhancement terms. Certainly we cannot overlook the "use" and "used for" terms, which are also known as "see" and "see also" as well as cross-references. Perhaps an attempt to discriminate among the terms would be usef ul.

Subject headings have long been associated with book cataloging and card catalogs. With the advent of computers and database searching in journals and periodicals we have seen the introduction of key words and descriptors as indexing aids.

A "thesaurus" is a compilation of words and phrases showing synonymous, hierarchical and other relationships. The function of a thesaurus is to provide a standardized vocabulary for information storage and retrieval. To accomplish this, use is made of bro ader terms, narrower terms and related terms as well as "use" and "used for" terms. A thesaurus is more closely identifiable with subject-authority lists, although, strictly speaking, an authority list is designed to establish the correct forms and spelli ng of a heading in an effort to promote consistency and prevent duplication of a subject term under different headings. To accomplish this cross-references, "see" and "see also," are introduced as required.

An "enhancement term" is used, typically in full-text online databases, to augment or enhance a story or article where such a term is required to search effectively for the story subsequently. Such terms are selected from a growing list of approved enhanc ement terms.

An "index term," either a single word or a phrase, is associated with a thesaurus to represent a single concept. Some might refer to it as a "topic." Almost synonymous with "index term" is "descriptor," which is referred to as an element of a controlled v ocabulary, for example, a thesaurus. In online search and retrieval strategies, logical and proximity operators are relied upon to increase the precision of information retrieval. Printed indexes, since they are precoordinated, do not have this degree of flexibility. To soften the impact of this limitation, wise and experienced indexers keep their thesauri dynamic.

"Identifiers" are key words usually extracted from the text of a document to represent a specific aspect of its contents. Occasionally, an "identifier" is referred to as a "junior descriptor" since it does not come from a controlled vocabulary list or the saurus. An "identifier" may become, in time, a "descriptor."

A "key word," or "keyword," as it is sometimes spelled, is described as a primary content-bearing word.

Finally, for our purposes, we agree that a subject-authority file (or a controlled vocabulary list or a thesaurus or list of approved topics or selected key words) is essential to an indexing project because it maintains consistency and accuracy in headin gs. In addition to subject headings or indexing terms, the authority file should include a network of cross-references to guide users from unselected terms to preferred ones (called "see references") and from usable terms to other related ones (called "se e also references"). This effort reduces the inadvertent selection of synonyms as headings for articles relating to a single concept. The following are examples of "see" references:

  • AIDS
    see Acquired Immune Deficiency Syndrome
  • Larceny
    see Robbery
  • C.E.T.A.
    see Comprehensive Employment and Training Act

To guide indexers and users to related subject terms headings, a subject authority file with "see also" references is essential. The following are examples of "see also" references:

  • Acquired Immune Deficiency Syndrome
    see also HIV
  • Robbery, Durham, NC
    see also Armed Robbery, Durham, NC
    see also Bank Robbery, Durham, NC
    see also Crimes, Durham, NC
  • Comprehensive Employment and Training Act
    see also Employment, Durham
  • Sleaze Factor
    see also Political Corruption

A subject authority list also includes scope notes in parentheses to clarify ambiguous terms. The following are examples of scope notes:

  • The Women's Room (book)
  • Pizza Hut (Erwin Rd.)
  • Oakwood (housing subdivision)
  • Oakwood (restaurant)

Whether the authority file is compiled before or during the indexing process, it will continue to grow. New places, organizations and events constantly appear in the news, so provision for continuous growth should be made during the planning stages.

Some indexers maintain the authority file on catalog cards or typed lists. Some rely on computer printouts. Still others consult online versions. All of these varieties can be expanded, but the catalog-card and typed-list systems have the disadvantage of providing only one copy, which must remain in one place to be used by one person at a time. They also consume more physical space and are prone to errors caused by misfiling. The printout and online forms, while more convenient and easier to update and co rrect, can be more expensive.

Specificity of Headings

When deciding whether to use general or specific subject headings, the most important consideration is the user. An indexer should reflect on how people are likely to ask for information and index accordingly. Direct, specific headings are easier for use rs to find because they expedite the location of specific items.

Newspaper indexing does not lend itself well to hierarchical arrangement of headings. The dictionary approach is more effective with users. The challenge with specific headings is to avoid trouble with synonyms, primarily through cross-references and scop e notes, and the scattering of terms. The latter problem can be mitigated with copious and well-controlled cross-references and standardized sub-headings. Subject-heading policies for the selection and format of new terms must be clear.

Rarely mentioned, but of crucial importance to a quality index, is making certain that similar or related articles appearing over a period of months or years can be captured by a single "master heading." Such a designated heading would be selected consist ently by indexers for a variety of topics requiring broad as well as narrow access. This approach is not a seeming reversal of the specificity principle in newspaper indexing. It is an attempt to reduce "scatter" in retrieval by helping users locate all r elated articles on a broad topic easily.

Nazir Ahmad gives effective testimony to the need for comprehensiveness, consistency and specificity in newspaper indexing.54 Quality control in newspaper indexes depends upon the effectiveness of subject analysis of news items. His article describes prac tical aspects of conceptual analysis and offers reasons for providing multiple subject entries for news stories. Ahmad lists rules for news analysis and for accurate allocation of subject headings. The article is enhanced with examples of nineteenth- and twentieth-century printed international newspaper indexes.

Indexing Standards

Although newspaper indexes vary considerably in purpose, scope, format, content and vocabulary, there are published standards to help the neophyte newspaper indexer understand the complexities of the indexing process. Recommended for study are the Americ an National Standard Institute's publications Z39.4, titled Basic Criteria For Indexes, and ANSI Z39.19, titled Guidelines for Thesaurus Structure, Construction and Use. Both are available at reasonable cost from the American National Standards Institute, Inc., 1430 Broadway, New York, NY 10018.

From abroad we learn of the British Library Newspaper Library's efforts in not only listing newspaper indexes (700 of them by 1987) but in organizing seminars on newspaper indexing standards.55 Several international newspaper indexing projects, historical and current, involving standards, policies and guidelines are described in detail by Peter Johnson,56 Nazir Ahmad,57 Geraldine Beare,58 Robert Stafford59 and E. E. Okorafor.60

Betty Eddison and David Batty's article on database design deals thoughtfully with the need for standardized subject terms.61 It offers effective approaches to thesaurus construction. Although the article is directed toward online databases, it has consid erable value for the nonwired indexer as well. Its succinct history of index language development is a gem. Although not a standards guide, Trudi Bellardo self-study workbook offers an excellent opportunity to learn the basic concepts of indexes and speci al subjects.62

Indexing Procedures

Along with the written guidelines for indexing articles, there should be a written series of procedures for indexers to follow when working with the articles and preparing the index. In order to select the articles to be indexed, the indexer must careful ly search through the newspaper. It is vitally important to check which edition of the newspaper is being indexed. Be certain that it is the one(s) of record, for example, the microfilmed edition(s). Local bylines and datelines are often clues indicating material of local interest. Once an article has been chosen for inclusion in an index, it is designated in some way, such as outlining the article, placing a check mark near the headline, underlining key words or names or writing subject terms below the h eadline.

When one is indexing a news article, names or indexing terms are generally underlines if they appear in the article. If appropriate headings do not appear in the text, they may be written in. At some point, headings should be checked against the subject t hesaurus to be certain that word order, spelling and terms are correct and consistent with established practice. If new subject headings are being established, reference tools such as the local city directory, gazetteer, telephone book or a reputable dict ionary and encyclopedia should be consulted for verification of names, places and spellings. Consistency in assigning subject headings to infrequent, but continuing, stories is essential to the production of a quality index.

Some indexers copy the information from each article onto a standardized form that is then used in compiling the index. This method, however, creates more opportunity for error and is so time-consuming that the practice of working directly from the marked newspaper is recommended. This method is not an option when doing retrospective indexing on newspapers that are bound or microfilmed.

The actual step-by-step process for producing an index depends upon the computer system and the software designed for it. Once the information has been entered into the computer, editing (comparing the marked newspaper with the article entry printout or w ith a form for microfilm indexing) eliminates inaccuracies and identifies omissions. Daily editing is more efficient than editing a large mass of material at the end of a month.

Index Staffing

Qualified personnel are crucial to starting and maintaining quality indexing projects. Depending upon the volume of material to be indexed and the method of indexing, the number of employees needed may vary. For small projects, one person to index, type and proofread might be adequate. On larger projects, a division of labor between two or more people is preferable. This allows the advantage of ensuring adequate staff backup during vacations and other absences to avoid delays in production. Further, it p romotes a broader staff familiarity with indexing policies and procedures.

Indexing is not just a mechanical function but one that involves imaginative and critical judgments. It is recommended that professional librarians or indexers do the actual indexing because of their training in managing information storage and retrieval operations. Part-time employees or dedicated volunteers enlisted to do the typing or data entry can relieve the indexer of many of the clerical functions associated with producing an index. The tasks that cannot be neglected by the professional indexer ar e editing and proofreading. To do so is to endanger the quality of the index.

Filing Rules

It is almost impossible to write anything brief about specific filing rules. They are, for the most part, a collection of arbitrary practices governing filing order. The hard part is keeping track of the decisions affecting the exceptions to the filing r ules. In truth, filing rules have frustrated and challenged even experienced librarians.

Before 1980 American Library Association filing rules were based upon "meanings" rather than their graphic representation. In other words, filing arrangement was based upon the type of entity represented in a heading or the form of a name whenever multipl e headings begin with the same word. In short, meaning and form took precedence over alphabetical order. Development and increasing reliance upon electronic files diminished, but did not solve, this problem.

Since 1980 these file-by-meaning rules have been abandoned by most libraries and librarians in favor of the "file-as-is" rules. ALA's new filing rules were reduced to a 50-page pamphlet containing ten main rules and a few subrules. The consideration for f iling order now is based upon the graphic representation of words, numbers and symbols. This advancement does not eliminate the further need for indexers to make filing decisions, but it does simplify the process. For the best treatment of these few, but critical, details, consult Hans Wellisch's section on filing in his 1991 book.63 Another fine source is Ruthanne Lowe's article.64

For the newspaper indexer, it comes down to selecting a general filing-order guideline to follow and making well-documented exceptions. The guiding principle should be to sort according to word-by-word order, but file as is. This means that people, places , subjects and titles with the same name are filed alphabetically, unless the index is divided into two major sections, personal names and topics or subjects. In that case, the same filing rule prevails, but in two indexes instead of one. Other examples i nclude the following:

Mc/Mac/M' are no longer lumped together in a single a single alphabet. They, also, file as is.
Numbers file as is. Numbers expressed as numerals (arabic or roman) precede words consisting of letters and are arranged according to their numerical value. Numbers expressed as words are filed alphabetically.
Abbreviations are arranged as written.
Inverted headings interfile with qualified headings.

Correction Policies

A characteristic of a published index is that once it is made, it is permanent and not subject to change. Those errors noted after publication must wait weeks or months until an update is scheduled before a correction may appear. Newspaper clippings, on the other hand, can and should have corrections, clarifications and even the rare retractions appended directly on the errant article. The same policy holds with online full-text news databases. It is part of the quality control system for the clipping fi les and the databases.

Extreme care in this area is taken by newspaper librarians managing clip files. If more than one copy of an article is filed under separate headings, each copy under all those headings must be targeted for the added correction or clarification. The correc tion procedure varies slightly with the online news databases. There is usually a correction field in online article records. Should a correction be published in the print edition, it is noted in the correction field. Management of unpublished corrections is usually addressed by individual database providers.

These careful methods make as certain as possible that errors are not repeated. A commitment to accuracy and excellence will also help avoid legal pitfalls. To be credible, printed news indexes must reflect this same high standard regarding corrections.

Public Access Policies

News libraries associated with midsized to large metropolitan papers usually restrict full access to their collections and services to in-house staff. This general policy is understandable given limited library staffing and the never-ending demand for da ta on deadline from reporters, editors and photographers. Nevertheless, this issue has newspaper libraries divided. For a discussion of varying news library positions, consult former St. Petersburg Times librarian James Scofield's chapter on public servic e65 and former Windsor Star librarian Mary Jane Handy's survey on public access in selected Canadian newspaper libraries.66

A different approach seems to apply for small-circulation dailies or community papers. As a rule, these papers do not have libraries, but are often willing to make an effort to accommodate the occasional request from their readers for access to bound volu mes, tearsheet files or boxes of local photos.

Some newspaper libraries have evolved a limited public-access policy. While restricting drop-in searchers, they will provide limited service by telephone or mail. These public requests must be relatively simple, requiring only a quick check for a date of a recently published article or photo to enable the inquirer to purchase a back issue of the paper or check the local public library's newspaper microfilm. Rarely do any newspaper libraries permit direct public access to clipping files.

To cope with these problems limiting public access to local news, public and academic libraries for decades have selectively clipped articles from local newspapers for their vertical files. Such files are admittedly incomplete and ephemeral, with inadequa te subject-heading control. Perhaps the frustration with this operation, by both librarians and the public, has prompted the move toward indexing local newspapers.

The availability of a quality local newspaper index greatly relieves news library staffs from coping with public inquiries. Even if news librarians themselves do not undertake such indexing projects, they should recognize the widespread benefits and find ways to encourage corporate support for news indexing efforts by public and academic libraries. There are other values of indexes to newspapers that will be mentioned later in this chapter.

Marketing the Newspaper Index

While reference librarians in public and academic libraries are quick to recognize the advantages of an index to local newspapers, they are perhaps less acquainted with the financial realities of supporting such an undertaking. In truth, there is no real market for local newspaper indexes beyond their immediate geographical area. Susan Cherry in 1979 reported that no matter how much people say they want local newspaper indexes, they do not want to pay for them.67 She described five public libraries (in F lint, Michigan; Minneapolis and St. Paul, Minnesota; St. Louis, Missouri; and Cleveland, Ohio) that spent hundreds of hours indexing local papers in various formats. These were money-losing propositions. Of course, the libraries did not expect to make a p rofit, but it was obvious that they were not even going to break even financially. In fact, the projects had to be heavily subsidized to continue operating. For example, in 1979 annual subscriptions to the St. Paul Public Library's index to two local pape rs cost $120 for libraries and public-supported groups and $160 for private organizations. At that time the library had 28 subscribers generating a maximum annual income of $4,480, but the indexing operation cost the library $35,000 to produce each year. Nevertheless, the index was considered a public service, and no thought was given at that time to eliminating it or even reducing coverage so as to lower costs.68

Another more recent example is the index produced from 1981 to 1990 by The Herald-Sun newspapers in Durham, North Carolina.69 Operating costs for maintaining the computers and software, salaries of two to three indexers, paper and supplies, printing and d istribution expenses ran an estimated $55,000 per year. Subscriptions to the monthly index with annual cumulations started at $150 per year, increased to $300 by the end of the nine-year project and never exceeded a total of 15 even with active promotiona l efforts. Marketing information searches for a fee on the computer index yielded only an additional $500 to $600 per year. The annual revenue generated from this project did not even reach 10 percent of the costs. Beyond the service value, it would be di fficult for a small newspaper to continue to bear full financial responsibility for an extended, indefinite period of time.

Value of Newspaper Indexes

Newspaper indexes have value beyond the subscription revenue. Newspapers with indexes, particularly good ones, discover that there is greater demand for the microfilm of their editions. Newspapers get royalties from microfilm vendors, usually a percentag e of the sales. In the case of The Herald-Sun newspapers, sales of subscriptions to the microfilm increased from 4 to 13 as a direct result of the availability of the index.

Indexes have further value for local papers. Organizations interested in soliciting some financial support from newspapers for such a project should seek to capitalize on these special needs. For example, an index would be valuable to a newspaper if it co uld supply the dates of important articles missing from a clip file or assist with the sale of back issues by providing dates of articles or photos and graphics published in the paper. Depending upon the scope and depth of local indexing, an index could a ssist in quickly identifying unique articles, features, profiles or series covering hazardous-waste treatment sites, bond issues, congressional redistricting or other complex topics.

Desmond stated repeatedly that a newspaper may reduce the material in its clipping file by as much as 30 percent with an index.70 In this situation, an index would not replace the clipping file (see the Bridget Janus's chapter in this book on clipping fil es), but would supplement it by indexing items of lesser matters unlikely to be referenced by the news staff except on rare occasions. A careful discussion of just what constitutes "lesser matters" is crucial. There might be surprising disagreement betwee n the news staff and public librarians on this topic. How much a local newspaper would pay to support such an indexing operation is a matter of individual negotiation.

Of course, many libraries attempt to undertake local news indexing in spite of severely restricted resources. They use library volunteers or add indexing responsibilities to existing staff. Relying upon shared computers and printers helps to some degree. Seeking seed money from foundations can ease start-up costs with equipment and software purchases but does not help with ongoing maintenance expenses. Sharply limiting the scope of the index and the complexity of vocabulary control is a frequent method em ployed to reduce costs. Most of these methods work against the production of a long-term quality product. Consequently, local newspaper indexing projects rarely survive more than a few years. The good news is that they can be revived.

It is the very rare publication that can break even financially, let alone realize profits on the basis of circulation or subscriptions exclusively. Might not advertising be considered? A newspaper index with ads? Unthinkable? Perhaps. But on the other ha nd, why not?

Updating Indexing Operations

Early in the course of the indexing project, it is important to document all of these considerations about filing rules, article selection policies and subject-heading preferences. But what about those indexing operations in a state of transition? For ex ample, a project may be moving from a manual to a computer-assisted operation. Once started, are these early decisions unalterable? Of course not. However, change is never easy. There are always additional costs and delays. Some changes are even hazardous to your data. Back up your operating programs and your data before converting. Flexibility must be built into your indexing program. Plan for change.

The decisions made years ago may be inappropriate, ineffective or out-of-date. How can an interface between the two systems be created? Perhaps the article selection policy needs to be modified. Maybe the subject-authority policies need revision. Small de tails such as the number of spaces for the date code may become critical. In short, changes and improvements, unknown and unforeseen initially, will eventually be necessary.

Take time to visit other newspaper indexing operations and discuss policies and procedures, current and desired, with the indexers in charge. Spend some time actually using a proposed system. Carefully inspect printed indexes over a span of a decade. Note differences in content and format. Then arrange to do a sample index. Lesley Sandlin, June Schlessinger and Bernard Schlessinger's article offers some valuable insight into this practice.71 It effectively addresses the need for pilot testing when underta king the indexing of smaller circulation daily newspapers.

Chapter Notes:

  1. Hans H. Wellisch, Indexing from A to Z (New York: H. W. Wilson Company, 1991), 23.
  2. From the introduction in The Washington Post Index.
  3. Harry A. Friedman, Newspaper Indexing (Milwaukee, Wis.: Marquette University Press, 1942), 23.
  4. Henry B. Wheatley, How to Make an Index (London: E. Stock, 1902), ix-xii.
  5. Ibid., 10.
  6. Mary Lou Lathrop and Norman M. Lathrop, Lathrop Report on Newspaper Indexes (Wooster, Ohio: Norman Lathrop Enterprises, 1979), 1129A.
  7. Friedman, Newspaper Indexing, 8.
  8. Robert W. Desmond, Newspaper Reference Methods (Minneapolis: University of Minnesota Press, 1933), 1-9.
  9. Friedman, Newspaper Indexing, 9-10.
  10. Jeff Pemberton, "A Backward and Forward Look at the New York Times Information Bank: A Tale of Ironies Compounded . . . and an Analysis of the Mead Deal," Online 7 (July 1983): 7-17.
  11. Friedman, Newspaper Indexing, 12.
  12. Ibid.
  13. Ibid., 16.
  14. Ibid., 11.
  15. Agnes J. Petersen, "Questionnaire on Newspaper Libraries," Special Libraries 18 (December 1927): 325-331.
  16. Friedman, Newspaper Indexing, 17.
  17. Ibid., 8-11.
  18. Russell H. Zesky, "Newspapers on Microfilm: History As It Was Happening (And Indexes to Help You Find Your Way)," The Serials Librarian 4 (Summer 1980): 393-399.
  19. Barbara Newcombe, "On Newspapers and Indexing," ASI Newsletter (American Society of Indexers) 104 (May/June 1991): 1.
  20. Ibid., 25.
  21. Anne P. Mintz, "Quality Control and the Zen of Database Production," Online 14 (November 1990): 15-23.
  22. Colleen Daze, "Money-saving Index Saves Space, Too," News Library News 9 (Fall 1986): 1, 4-5.
  23. Gloria Colvin, "Bridging the Local Information Gap: A Proposal for Developing an Index to the Durham Morning Herald and the Durham Sun," master's thesis, School of Library Science, University of North Carolina at Chapel Hill, 1980.
  24. Lathrop and Lathrop, Lathrop Report.
  25. K. D. Sell, "Checklist of Published Indexes to Current American Daily Newspapers," RQ 17 (Fall 1977): 13-16.
  26. Anita Cheek Milner, Newspaper Indexes: A Location and Subject Guide for Researchers (Metuchen, N.J.: Scarecrow Press, 1977).
  27. Candace Peterson, "The San Antonio Express News: Indexing Considerations, or Have We Bitten Off More Than We Can Chew?" paper presented at the American Society of Indexers 24th Annual Meeting, May 22, 1992, San Antonio, Texas.
  28. Carol Cole, "The Care and Feeding of a Newspaper Index," paper presented to the Michigan Library Association in 1975.
  29. David Nelson, "Newspaper Indexing with Microcomputers: The Washington State Experience," Reference Services Review 17 (Winter 1989): 91-95.
  30. Linda Fetters. A Guide to Indexing Software, 4th. ed. (Philadelphia, Pa.: American Society of Indexers, 1992).
  31. Ted Hunsberger, "Using dBase III+ to Index the Local Newspapers," Library Software Review 8 (November/December 1989): 336-337.
  32. Michael D. Cramer and Mary J. Markland, "Newspaper Indexing with Pro-Cite," College and Research Libraries News 50 (October 1989): 807-808.
  33. Charles Anderson, "Getting Savvy: A Database for Reference Use," Wilson Library Bulletin (September 1986): 41-43.
  34. Michael Knee, "Newspaper Indexing Using WordStar and Personal Pearl," Reference Services Review 14 (Summer 1986): 31-32.
  35. Celia Wall, "Newspaper Indexing: Using an IBM Mainframe Computer and a Text-Formatting Program," Information Technology and Libraries 6 (March 1987): 34-39.
  36. Jean E. Koch, "Newspaper Indexing: Planning and Options," Special Libraries 76 (Fall 1985): 271-280.
  37. John Rothman, "Preserving the News That's Fit to Print," The Indexer 5 (Spring 1966): 39-42; and John Rothman, "The New York Times Information Bank," Special Libraries 63 (March 1972): 111-115.
  38. Alan R. Greengrass, "Indexing at the New York Times Information Service," in Indexing Specialized Formats and Subjects, ed. Hilda Feinberg (Metuchen, N.J.: Scarecrow Press, 1983), 180-188.
  39. Jessica L. Milstead, "Newspaper Indexing: The Official Washington Post Index," in Indexing Specialized Formats and Subjects, ed. Hilda Feinberg (Metuchen, N.J.: Scarecrow Press, 1983), 189-204.
  40. For a brief history of these indexes produced by Bell & Howell, consult Zesky, "Newspapers on Microfilm," 393-399.
  41. "A New Newspaper Index." The Indexer 18 (October 1992): 107. A sample page of The Independent Index is included with the announcement.
  42. Stephen Smith, "Online News Retrieval Systems: Evaluations and Library Applications," Reference Services Review 10 (Winter 1982): 47-60.
  43. Carol Tenopir, "Newspapers Online," Library Journal 109 (March 1, 1984): 452-453.
  44. Reva Basch, "DataTimes, VU/Text," RQ 25 (Summer 1986): 529-530.
  45. Dorothy L. Ingebretsen and Thomas M. Lutgen, "The Los Angeles Times: A Special Kind of Database," Database Searcher 7 (May 1991): 17-30.
  46. Mintz, "Quality Control and the Zen of Database Production," 15-23.
  47. Reva Basch, "The Seven Deadly Sins of Full-Text Searching," Database 12 (1989): 15-23.
  48. Automating the Newspaper Clipping Files: A Practical Guide (Washington, D.C.: News Division of the Special Libraries Association, 1987).
  49. Tefko Saracevic, "Indexing, Searching, and Relevance," in Indexing: The State of Our Knowledge and the State of Our Ignorance, ed. Bella Hass Weinberg (Medford, N.J.: Learned Information, Inc., 1988), 102-109.
  50. Louis M. Gomez, Carol C. Lochbaum and Thomas K. Landauer, "All the Right Words: Finding What You Want as a Function of Richness of Indexing Vocabulary," Journal of the American Society for Information Science 41 (December 1990): 547-559.
  51. Much of the material included in this section came from an indexing manual prepared by Gloria Colvin and Barbara Semonche in 1986.
  52. Bobbie Vierra and Tom Trice, "Local Newspaper Indexing: A Public Library Reports Its Experience," The Serials Librarian 5 (Fall 1980): 87-92.
  53. Thomas F. Armstrong and Janice C. Fennell, "Historical and Genealogical Gold Mine: An Index Project for a Small-Town Newspaper," RQ 22 (Winter 1982): 140-145.
  54. Nazir Ahmad, "Newspaper Indexing: An International Overview," The Indexer 17 (October 1991) : 257-266.
  55. Geraldine Beare, "Pressing for Press Indexing," The Indexer 16 (April 1989): 145-146.
  56. Peter Johnson, "Printed Indexes to Early British Periodicals," The Indexer 16 (April 1989): 147-155.
  57. Ahmad, "Newspaper Indexing: An International Overview," 257-266.
  58. Geraldine Beare, "Local Newspaper Indexing Projects and Products," The Indexer 16 (October 1989): 227-233.
  59. Robert Stafford, "Australia Newspaper Index Feasibility Study," Development of Resource Sharing Networks Study no. 13, National Library of Australia, Canberra, 1980.
  60. E. E. Okorafor, "Indexing Nigerian Newspapers for Effective Research," Information Development 5 (October 1989): 228-234.
  61. Betty Eddison and David Batty, "Database Design," Database 11 (December 1988): 109-112.
  62. Trudi Bellardo. Subject Indexing: An Introductory Guide. (Washington, D.C.: Special Libraries Association, 1991).
  63. Hans H. Wellisch, Indexing from A to Z, 130-144.
  64. Ruthanne Lowe, "In and Out of Sorts," ASI Newsletter (March/April 1988): 1, 15-17.
  65. James Scofield, "Public Service," in Guidelines for Newspaper Libraries 2d ed. (Reston, Va.: ANPA Foundation, 1983), 43-46.
  66. Mary Jane Handy, "Public Use of Canadian News Libraries," News Library News 7 (Spring 1985): 14-15.
  67. Susan Spaeth Cherry, "Yesterday's News for Tomorrow: A Special Update on News Indexes, Indexing, and Indexers," American Libraries 10 (November 1979): 590.
  68. Ibid.
  69. For a partial discussion of this project, consult Barbara Semonche, "Newspaper Indexing: In Search of a Solution," Collection Building 7 (1986): 24-28.
  70. Desmond, Newspaper Reference Methods, 118.
  71. Lesley McGee Sandlin, June H. Schlessinger and Bernard S. Schlessinger, "Indexing of Smaller-Circulation Daily Newspapers," The Indexer 14 (April 1985): 184-189.