College and Research Libraries WILLIAM J. MAHER and BENJAMIN .F. SHEARER Undergraduate Use Patterns of Newspapers on Microfilm Thi s analysis, based on unde rg raduate use patte rns of neu;spape rs on mi- crofilm at th e ( ni t.:e rsity of Illinois , prese nts criteria , Su ch as numbe r of ti- tles , dates requ ested , and existen ce of indexes, that should b e conside red u;hen purchasing microfilmed new s pape rs for th e highest possible benefit in te nns of mone y spe nt . Statistics on use suggest that a use r satisfa ction let;el of almost 85 pe rce nt can b e attained u;ith a trcenty-year backfile of ttce nty titles . Th e analysis presented is offered as one step in d et;elopme nt of an aCl(llis itions polic y for ne rcs pape rs on microfilm. IN RECENT YEARS newspapers have become increasingly important reference and re - search sources for ·a wide variety of users on college and university campuses . They pro- vide original documentation for events that are too recent to be covered in monographs and professional journals or that have been bypassed by the traditional literature of his- tory , political science , economics , and the other social sciences and humanities . News- papers are particularly attractive and useful to undergraduates for course papers and class projects on a variety of subjects. They do, however , present librarians with several unique problems . Because of their physical nature, storage and preservation are difficult . While microfilm does represent a solution , newspapers still present obsta- cles because of the high cost of microfilm runs and because of the lack of bibliographic control over their content. Moreover, since newspapers are general in their subject coverage , the determination of potential us e rs is difficult. Undoubtedly , these factors contribute to the lack of professional litera- ture on newspapers as library materials. 1 Few libraries have developed guidelines William ]. Maher is newspaper librarian , Uni- ve rsity of Illinois at Urbana-Champaign, and Benjamin F. Shearer is go vernment documents/ law librarian , East Tennessee State Uni versity, johnson City , Tennessee . 254 I for a newspaper acquisitions policy , and this article will attempt to establish criteria for the developme nt of a policy for newspapers on microfilm for undergraduat e users in medium-size and small four-year colleges and universities. It is based on an analysis of undergraduate use at the newspaper li- brary of the University of Illinois at Urbana-Champaign during the peak periods of the 1976177 academic year. While the conclusions presente d here may not hold true for all libraries, the methodology will be valid and should direct librarians to the formulation of policies for the selection and purchase of newspapers on microfilm. USER SURVEY Before presenting use statistics and guidelines for acquisition, it is useful to de- scribe the holdings and operations of the newspaper library at Illinois. With more than 63 ,000 re e ls of microfilm e d ne wspa- p e rs , th e library hold s about 324 U.S. and 262 foreign titles. While the collection is particularly strong in Illinois titles , there are large microfilm runs for metropolitan news- papers in at least twenty-five states and in fourteen European and fourteen non- European countries. There are strong hold- ings in such areas as labor , religious , liter- ary , and black newspapers . Because of an annual use of at least 20,000 reels of microfilm, 2 it is possible to analyze the exact nature of undergraduate interest in newspapers from statistics main- tained on their use. When requesting film, all users complete a reference card on which they record the desired title and dates as well as information regarding their purpose and status (i.e., faculty, graduate student, undergraduate, and public). These cards are used for all titles except the New York Times , the Chicago Tribune, and the Wall Street Journal. Because de- mand for these three titles is very high, they are stor~d in microfilm cabinets freely accessible to all patrons. All other titles are shelved in a restricted area , and patrons complete the reference card to request them . The cards provide the basis for the current analysis. While the three most popular titles-New York Times , Chicago Tribune , and Wall Street Journal-are not includ~d in these statistics, these titles are of such importance that they should be considered basic items for any college and university library. While the Chicago Tribune is heavily used at Illi- nois because of its local appeal, it is of na- tional importance and should be a priority for many libraries , especially those in the Midwest. (Another regional ·substitute might be the Washington Post .) Any library not holding the New York Times and the Wall Street Journal should regard these as the first titles to acquire on microfilm. The New York Times , in particular, should be a top priority because of its importance and be- cause it is indexed for its entire run. Be- yond these basic titles , acquisitions policy can be determined best by an analysis of undergraduate use. At the University of Illinois newspaper li- brary, two months~November and April- consistently account for the highest use of film for the period from 1973 to 1977. This reflects a heavy demand by undergraduates in order to complete their course work as each semester draws to a close. While undergraduates annually account for 37 per- cent of use, during the months of Novem- ber and April they account for 49 percent. This fact justifies the isolation of these months for the current study. The use dur- ing this period illustrates the importance undergraduates place on newspapers as re- Undergraduate Use Patte rns I 255 search materials. The purpose of under- graduate use is also instructive : During No- vember 1976 and April 1977, 85.5 percent of undergraduate use was for course pa- pers (the other· types of use recorded were " research for publication," " dissertation/ thesis research ," " classroom use, " and "other") . To identify those newspapers that will satisfy the most requests at the lowest cost, the following analysis studies titles re- quested , dates requested , and the impor- tance of an index for use. TITLES REQUESTED Since newspapers on microfilm represent large investments, libraries must seek some assurance that the money spent yields an acceptable cost-benefit ratio when com- pared to the level of satisfaction of user needs. The case in favor of such basic titles as the New York Times , Wall Street jour- nal , and Chicago Tribune is clear, but the case for purchasing other newspaper titles is far more complex. The geographical factor is perhaps the single most obvious element that influences undergraduate use of newspapers on mi- crofilm. Thus the statistics presented here on titles requested at Illinois should not be taken as an inflexible acquisitions policy statement for other libraries . Some general principles on undergraduate use patterns do, however, emerge from these data, and these can be applied to any undergraduate library in terms of both titles and dates that may offer the greatest benefits for money spent. Table 1 shows user demand according to titles requested during November 1976 and April 1977. As this table indicates, under- graduate use of newspapers on microfilm during the two peak use months of the year was restricted to forty-nine titles (excluding the New York Times , Wall Street Journal, and Chicago Tribune ). Of the titles re- quested, thirteen (26.5 percent) were titles no longer published (" dead"), but in terms of total use these titles accounted for only sixteen requests (5 . 7 percent). The over- whelming number of requests were then for currently published ("live") titles. The level of user satisfaction increases as more titles are added. The data show that 256 I College & Research Libraries • May 1979 TABLE 1 NEWSPAPER TITLES ON MICROFILM REQ UESTED BY UNDERGRADUATES , NOVEMBER 1976 AND APRIL 1977, BY NUMBER OF REQ UESTS Number of Title Requests 1. Dailh Illini 31 2. Was ington Post 31 3. Morning Courier (Champaign) 30 4. Los Angeles Times 22 5. News-Gazette (Champaign) 19 6. St . Louis Post-Dispatch 13 7. Chicago Daily News 11 8. Times (London) 11 9. San Francisco Chronicle 8 10. Chicago Defender 7 11 . New York Amsterdam News 7 12. Atlanta Constitution 7 13. Chicago ·Sun-Times 6 14. Christian Science Monitor 6 15. Pittsburgh Courier 6 16. Village Voice 5 17. Los Angeles Sentinel 4 18. Detroit Free Press 4 19. Illinois State Journal- Register 4 20 . Louisville Courier-Journal 4 21. New York Herald Tribune (Dead) 4 22. Cleveland Plain Dealer 3 23. Dallas Morning News 3 24 . National Observer 3 25. Norfolk Journal & Guide 3 26. Cincinnati Enquirer 2 27 . Michigan Chronicle 2 28 . New Orleans Times-Picafune 2 29-49. Twenz-one additiona titles (12 " ead") with 1 request each 21 Total requests 279 47.7 percent of all requests could be satisfied with the top five titles, 65.6 per- cent with the top ten titles, 77.1 percent with the top fifteen , and 84.6 percent with the top twenty. Thus, only twenty-seven requests (9. 7 percent) for "live" titles were not satisfied by the twenty most popular ti- tles. The factor of local bias is indeed quite important, as these statistics show. Three of the top five titles requested are local news- papers and by themselves constitute 28.7 percent of all requests. The Daily Illini is the student paper of the University of Illi- nois, and its heavy use is quite understand- able. The Champaign-Urbana Cqurier and th.e News-Gazette are the two local papers. Language and country of origin are also sig- nificant factors. All of the top twenty titles requested w~re English-language titles, and the London Times was the only foreign title in the top twenty. Black newspapers also received relatively high usage-four of the top twenty were black papers (Chicago De- fender , New York Amsterdam News, Pittsburgh Courier, and Los Angeles Sen- tinel). The statistics on titles requested point to some important conclusions. A high level of user satisfaction can be attained with rela- tively few newspaper titles on microfilm. Aside from the school newspaper, six titles (including the two local papers; two titles from nearby metropolitan areas, such as the St. Louis Post-Dispatch and Chicago Sun- Times ; and two titles of national promi- nence, e.g., the Washington Post and the Los Angeles Times) could well satisfy 50 percent of all requests. It is necessary, therefore, that libraries consider geograph- ical bias, country of origin, language, and special topics taught in the curricula (e.g., black studies) when deciding to purchase microfilmed newspaper titles to meet user demands as precisely as possible. DATES REQUESTED Consideration should also be given to the dates requested. An investment in large backfiles of e~en frequently used titles may not prove beneficial if the use of certain older dates is minimal. Table 2 demonstrates that undergraduate requests for newspapers in microform de- crease significantly for the years preceding 1969. In fact, 19.7 percent of all requests could be satisfied by keeping only a one- year backfile in paper, thus avoiding a con- siderable expenditure for microfilm. This saving must, however, be weighed against the overall impracticality of storing and ser- vicing newsprint. Problems of theft, mutila- tion, and preservation must be taken into account. Assuming that libraries interested in building a research collection will consider purchasing all but current issues of news- papers in microform, ·it is instructive to analyze use in terms of five-, ten-, fifteen-, and twenty-year backfiles. It should be noted first, however, that only 22.9 percent of all citations were for dates before 1962, TABLE 2 DATES OF UNDERGRADUATE REC) UESTS FOR NEWSPAPERS ON MI CROF ILM , NOVEMBER 1976 AND APRIL 1977 umber of Year Requests 1976 117 1975 71 1974 24 1973 40 1972 33 Total 1972-76 285 (47.9%) 1971 34 1970 28 1969 27 1968 13 1967 13 Total 1967- 71 115 (19.3%) Total 1967-76 400 (67.2%) 1966 15 1965 8 1964 8 1963 19 1962 9 Total 1962-66 59 (10.0%) Total 1962-76 459 (77.1%) 1961 7 1960 4 1959 5 1958 21 1957 7 -- Total 1957-61 44 (7.4%) Total 1957-76 503 (84 .5%) 1947-56 40 1937-46 16 1927--36 1 1917-26 6 1907-16 4 1897-1906 11 1887-96 5 1877-86 1 1867-76 2 1857-66 6 Total 1857-1956 92 (15.5%) Total 1857-1976 595 (100%) and 8. 7 percent for dates before 1946. These are significant percentages of all cita- tions, but the cost of filling user demands for citations prior to 1962 must be deter- mined not only in r e lation to b e nefits but also in regard to the availability of older papers on some kind of cooperative loan arrangement. Interlibrary loan and membership in sys- tems, such as the Center for Research Li- braries, are good examples of cooperative arrangements. Certainly most "dead" titles , Undergraduate Use Patterns I 257 which are seldom used , are available through cooperative lending. In short, there is little reason for a four-year college library to invest in long backfiles of newspapers on microfilm, especially if budgets are tight. If a library is experiencing an especially severe budget crunch, it may choose to let its collection of newspapers in microform grow organically by ordering only current subscriptions. Th e cost of current subscrip- tions of filmed newspape rs usually is sig- nificantl y less than the cost of backfiles. For instance, based on the 1978 catalog of the Microfilming Corporation of America, a cur- rent subscription to the Atlanta Constitution costs $362, whereas r ece nt backfil es cost $540 per yea r. In general, recent backfiles (within the past fi\'e years ) tend to be about 40 percent more expensi\'e than the current subscriptions themseh-es. Although current subscriptions to approx- imate ly six titles may satisfy as much as 20 perce nt of user demands , more user satis- hlction is attained as the size of the backfile increases. Thus, with a fi\'e- yea r backfile (1972-76), 47 .9 percent of all requests could b e fill e d. With the purchase of a ten- yea r backfile (1967-76), 67.2 perce nt of all user d e mands could be satisfied; with a fifteen - year backfile (1962-76), 77.1 p e rcent of de- mands ; and with a twent y-yea r backfil e (1957-76), 84.5 percen t of all demands. These figures illustrate the types of ques- tions librarians must ask themseh es when deciding to purchase backfiles of mi- crofilmed newspapers. For example , if 77 . 1 p e rce nt of all requests can be satisfied with a fifteen- yea r backfil e, is the purchase of fi\'e additional years of a certain number of titles worth satisfying an additional 7.4 per- cent of user demands ? These statistics indi- cate that a fifteen-year backfile in microform of th e most fi·equentl y used newspapers (as dete rmin ed by the factors described abo, e ) oflers a , ·ery high benefit f(.>r its cost. The money that could be spent on fi, ·e more yea rs of backfiles might be better used to purchase new titles . When the data on titles requ es ted and dates r e qu es ted are brought together, a sig- nificant parallel emerges in the le\'el of user satisfaction that can be attained with num- bers of titles and the extent of backfile s. Summaries of user satisfaction b y number of ~------------------------------------------------------------------------------- 258 I College & Research Libraries • May 1979 titles and size of backfile are shown in tables 3 and 4. The close relationship shown in tables 3 and 4 between titles and years requested for user satisfaction demonstrates the necessity of considering these two factors together when deciding to purchase newspapers in microform. For example, a 77 .1-percent level of user satisfaction could not be at- tained merely by acquiring fifteen titles without also acquiring a fifteen-year backfile of these titles. The two criteria of number of titles and years of backfiles are inextricably related and must be considered jointly if a library wishes to obtain the highest level of user satisfaction for the money it spends on ' newspapers in microform . TABLE 3 USE-H SATISFACTJO!\ BY TITLES HELD l'erct•nt nf Ht'((llt'SI' An"''t"red 47.7o/c 65 .6 o/c 77.1% 84.6o/c TABLE 4 · umhPr nr Title, 5 10 15 20 USEH SATISFACTJO!\ BY YEAHS R E<)l'ESTED J'erl"Pill nf Ht·qnt"'h Answt•red 47.9o/c 67 .2% 77.1% 84.5o/c Yt·ar' or Baddilt• 5 (1972-76) 10 (1967-76) 15 (1962-76) 20 (1957-76) ROLE OF INDEXES Date and number of titles are important considerations in acquisitions policy, but ac- cessibility is equally significant . Newspapers are a major source of information on numerous subjects of great interest to undergraduates . They do, however, present problems for bibliographical control because they are comprehensive in their subject coverage. There are no indexes covering all news- papers, with the exception of microform se- ries, such as Newsbank, which contain top- ically arranged clippings fi·om several news- papers. While such collections do provide subject access to newspapers , their coverage is not deep and the subject categories are not likely to be of long-term interest to serious researchers. Moreover, their cost (about $1,000 per year for Newsbank) may seem high for the benefit they bring. 3 Thus adequate access to newspaper ac- counts on a given subject can be obtained only by the use of the available indexes for individual titles. Even then, the researcher is still confronted with a serious problem because of the lack of indexes for all but the most recent years of most newspapers. While the New York Times and the London Times are indexed for their entire runs, most metropolitan U.S. newspape rs were not indexed until about 1970. It is reasonable to argue that the exis- tence of an index will contribute to more requests for a given title, and statistics at the UniYersity of Illinois support this con- clusion. Table 5 describes the relation of use of indexed papers to use of papers not indexed . These statistics lead to seYeral important conclusions. First, for dates requested within the past ten years, indexed titles ac- count for more than one-half of all newspa- pers used. The average beginning date of most newspaper indexes is 1970 and is cal- culated on the following titles: Atlanta Con- stitution (index begins 1971), Index tv Black Newspapers ( 1977 ), Christian Science TABLE 5 ll\DEXED N E\\'SI'AI'EHS ' s II AHE OF TOTAL U~E 1976 1975 1974 1973 1972 1972-76 average 1971 1970 1969 1968 1967 1967- 71 average 1967-76 average 1966 1965 1964 1963 1962 1962-66 average 1962-76 average J't'l't'('lll nf U"· 78.6% 59.3% 62.5% 60.9% 82.6% 68. 1o/c 39.1% 52.4o/c 36.0o/c 50.0% 57.1 o/c 46.9% 51.8o/c 50.0 o/c 42.9% 50.0% 5.3% 1.0% 29.8% 48.5% Monitor (1950), Daily Illini (1961), journal of Commerce (1976), Los Angeles Times (1970), Minneapolis Star and Tribune (1971), National Obsen;er (1969), St. Louis ·newspapers (1975), San Francisco Chronicle (1970), New Orleans Times-Picayune (1972), and Washington Post (1971). The New York Times, Times of London, and Wall Street journal were excluded from the average. Thus it would be more appropriate to use the 1972-76 average (68. 7 percent) as indic- ative of undergraduate preference for in- dexed titles. In view of this, formulators of acquisitions policy should be very cautious about the purchase of newspapers on microfilm unless an index is available and also purchased. While the cost of indexes is high (averaging about $185, or one-half of the cost of many film subscriptions ), they clearly contribute to a large part of undergraduate use of newspapers in microform. 4 To a large extent, the purchase of an index for a title already held is preferable to acquiring new titles. For example: If a li- brary were receiving the Washington Post and the St. Louis Post-Dispatch at a total cost of $876 and were considering the pur- chase of another title such as the Times- Picayune at a cost of $390, it would be bet- ter advised to use that money for the in- dexes to the Post and the Post-Dispatch. Thus the selection of titles would be more limited, but access would be simplified and user satishlCtion could be increased . If a li- brary cannot bear the cost of monthly and/or quarterly indexes, it may consider purchasing only annual cumulative updates when these are available separately. It is logical to assume that overall use of newspapers will increase with the presence of indexes. There is, however, no statisti- cally valid trend in the increase of use with the availability of more indexes. Any ap- pearance of an increase in use with the presence of more indexes since 1970 must be tempered by the fact that more recent years are generally more desirable (as illus- trated in table 2). For an accurate appraisal of the eflect of indexes on overall use, it will be necessary to wait an additional five years when the use for the period 1972-76 is not influenced so heavily by its recency. While it is not Undergraduate Use Patterns I 259 possible to show an overall rise in use be- cause of indexes, there is a demonstrable increase in the preference for indexed newspapers. This would seem to be a natu- ral reflection. of, the greater availability of indexes since 1970. Considering the importance of indexes for the use of newspaper titles, one would be inclined to purchase only indexed titles. In- dexing should not, however, be the sole cri- terion because certain titles (e. g., those fi·om many major U . S. cities) will be heavily used despite the lack of an index. Of the twenty most requested titles listed in table 1, only eight (40 percent) were in- dexed at the time of use (four of the titles were black newspapers that have since been covered by the Index to Black Newspapers ). The fact that 60 percent of these twenty most-requested titles were not indexed should cause librarians to think carefi.tlly he- fore refusing to purchase any nonindexed ti- tles. Of the remaining twelve titles, five (two from Champaign-Urbana, two from Chicago, and one fi·om Springfield, Illinois) are "locally" focused papers that have no index . If these five "local" newspapers were excluded fi·om the analysis of the relation of indexes to most-used titles, indexed papers would account for 53.5 percent of the fifteen most-requested papers. It should also he emphasized that 68.7 percent of all use of newspapers covering the years 1972-76 is in indexed titles, even if indexed papers account for only 40 per- cent of the twenty most-requested titles. Thus if a library were contemplating the purchase of two papers of equal stature, one indexed and the other nonindexed, the in- dexed title should be preferred to achieve greater patron satishlCtion. More