College and Research Libraries Enhancing CD-ROM Searches with Online Updates: An Examination of End-User Needs, Strategies, and Problems Susan K. Charles and Katharine E. Clark Many academic researchers are interested in obtaining the most current information available when conducting computerized literature searches. A research project was conducted to evalu- ate the feasibility of providing an online updating service to end-users searching CD-ROM databases. Searchers using Silver Platter's Agricola database were asked to participate in the study. End-user searching strategies were observed and problems were noted. Factors such as cost-effectiveness and convenience were also studied. Data and observations indicate that end- users are interested in the availability of a low-cost, hybridized CD-ROM/ online updating sys- tem. Conclusions are drawn concerning the future direction of online updating systems for end-users. t a growing number of research institutions, CD-ROM data- bases are becoming increas- ingly important, high-profile tools for conducting literature searches and represent the trend to streamline end- user search technologies. The high-usage levels of these products at Texas A&M University's Evans Library is evidence of their popularity with a substantial seg- ment of library patrons. CD-ROM products offer patrons the op- portunity to obtain information through inexpensive, user-friendly formats, but CD-ROM databases are not as current as their online counterparts. In CD-ROM database updating, manufacturing and marketing procedures such as file master- ing, shipping, and processing cause de- lays that may be significant. 1 Enhance- ment of these databases with online services can provide searchers with access to more current data. Online services and CD-ROM products can be coordinated to create a search environment in which pa- trons are provided with the most appro- priate service. 2 Creating a system where end-users can complement CD-ROM database searches with current citations from online updates may be a practical way of providing patrons with a more complete service option. Current library literature reveals no evidence of studies that examine the use of an online end-user system to enhance a database offered as a CD-ROM product, despite the existence of systems produced by DIALOG, Wilson, and Lotus. These systems allow patrons to switch conveniently from a CD- ROM database to the online counterpart. Produced by Silver Platter, Agricola on CD-ROM is one of the most heavily used databases in the Reference Division of the Evans Library. This CD-ROM database is Susan K. Charles is an Infonnation Analyst for Hewlett-Packard Company Library MS 71, 3404 East Harmony Road, Ft. Collins, Colorado 80525-9599; and Katharine E. Clark is the Assistant Head of the Reference Division at the Sterling C. Evans Library of Texas A&M University, College Station, Texas 77843-5000. 321 322 College & Research Libraries updated quarterly and patrons who utilize this service are not accessing the latest six to nine months of agricultural literature unless they supplement their CD-ROM searches with alternative options. Patrons can access Agricola online file updates by paying a slight fee ($2.50 or more) to use the After Dark end-user service, available in the evenings and on Saturdays. Patrons can also access Agricola online by schedul- ing a mediated online search. All patrons, except those graduate students who are entitled to receive one free mediated search each academic year, must pay a fee to cover connect, telecommunications, and citation charges for mediated searches. Because of its popularity, Agri- cola was an ideal target for a study that ex- amined the use of an online updating sys- tem to complement end-user CD-ROM database searches. AVAILABILITY OF END-USER SERVICES AT THE EVANS LIBRARY At Texas A&M University, the availabil- ity of a variety of low-cost or free end-user services could affect the level of end-user expertise and, consequently, the success of any particular strategy to coordinate the use of complementary online and CD- ROM databases. End-users have access to twenty-eight different laserdisk databases that utilize a variety of search techniques and command structures. Sixteen of these databases are located in the Wiley Laser- disk Service Area of the Reference Divi- sion and the remainder are available in other public service areas such as the Doc- uments Division and Microtext Depart- ment. During a one-week period in April 1988, surveys collected in the Wiley Area indicated that 727 end-users had spent an average of 31 minutes using laserdisk databases. Agricola on CD-ROM ranked fourth highest in number of users. Ap- proximately 70% of the laserdisk users were undergraduates. 3 Another popular end-user service is the After Dark Service that enables students, faculty, and staff members to access ap- proximately 150 databases through the BRS/ After Dark, Knowledge Index, and STN systems. A generous gift from the Association of Former Students helps July 1990 support connect costs for After Dark end- user searching by students, resulting in free access to all online databases available through Knowledge Index and substan- tially reduced charges for BRS/ After Dark. Subsidized end-user searching is also available through IHS's Tech Data Ser- vice. During the 1987-88 academic year, 2,384 online end-user searches were per- formed for a total of 1532.77 online hours. The average cost to the patron for all After Dark end-user searches was $.79. This fig- ure includes free Knowledge Index searches, but the average cost to the pa- tron for end-user searches that resulted in citation and/ or connect charges was still low at $4.68. Fifty-nine percent of end- users were graduate students, 35.5% were undergraduates, and 5.5% were Univer- sity faculty or staff. 4 These statistics are ev- idence that a substantial pool of Evans Li- brary patrons do access a wide range of cost-effective end-user services. OBJECTIVES The Agricola online updating study was conducted to meet several main objec- tives. First, the study was used to deter- mine if patrons searching Agricola on CD- ROM would be willing to use an online updating system to obtain more current ci- tations. Second, the study measured the level of satisfaction that these patrons ex- perienced after searching both Agricola on CD-ROM and the online updating sys- tem. Finally, an estimation of the cost of such a service to the Library needed to be determined if this type of system were to be made available to all end-users cur- rently searching Agricola on CD-ROM. Several types of "costs" were subse- quently evaluated. These included: (1) tangible, quantitative costs such as online search changes, and (2) staff time require- ments for coordinating this type of service and providing search assistance. METHODOLOGY A random sampling of CD-ROM Agri- cola users was conducted to select partici- pants for the online updating study. The study was conducted over a 13-week pe- riod in 1988. Each week was divided into 50 1-hour slots that were available for ran- ' . dom selection. A random number table was used to select those slots during which users of Agricola on CD-ROM would be offered an opportunity to up- date their search by using Agricola online at no charge. The first CD-ROM Agricola user to arrive during each randomized slot was approached about participating in the study. Patrons who chose to participate concluded their search on the CD-ROM database and then contacted the investi- gator when they were ready to do their on- line search. Patrons who declined to par- ticipate completed a brief form to explain why they were not interested in updating their searches online. After the search strategy was com- pleted, the participant went online, conducted the search, and selected online prints. Online Agricola searches were con- ducted using the DIALOG system and DIALOGLINK. communications software. The type ahead feature of DIALOGLINK. enables online users to type search state- ments offline. This feature was utilized for the updating study to allow participants to type and revise their searches at their own pace. Each participant was asked to read an instruction sheet that provided simpli- fied, step-by-step instructions on con- ducting the search and using DIALOG search commands. A brief sample search was also provided. After reading the in- structions, each participant began typing the search strategy previously employed for the CD-ROM search onto the appropri- ate DIALOGLINK screen. After the search strategy was completed, the partic- ipant went online, conducted the search, and selected online prints. All searches were limited to the most recent updates that included the last two months of cov- erage pr~d by the most current CD- ROM disk. Participants had a chance to examine the citations retrieved by their search during the printing process. Each session was monitored by one of the in- vestigators. Enhancing CD-ROM Searches 323 A questionnaire was used to measure and evaluate the responses of the partici- pants using the online system to update the CD-ROM searches. Participants pro- vided information on their status (under- graduate, graduate, faculty, or other) and their familiarity with computers, data- bases, and end-user services. Participants were also asked to evaluate the Agricola online updating system in terms of ease of use, time needed to complete the search, problems encountered, and the need for assistance. The number of citations ob- tained from both the online and CD-ROM searches were recorded, and participants evaluated the relevancy and usefulness of the citations retrieved by the online sys- tem. Participants gave the online updat- ing system an overall satisfaction rating and provided data on how frequently they would use the system if it were available on a permanent basis and what fees they would consider paying for the service . Many participants also provided addi- tional comments on the online updating system. STUDY RESULTS A total of 87 1-hour randomized slots were monitored for end-users willing to participate in the study. Out of the 45 peo- ple who used Agricola on CD-ROM dur- ing these time periods, 30 (67%) elected to update their searches using Agricola on- line. Of the 15 respondents who declined to utilize the online updating system, 8 stated that they did not have enough time to participate, 6 thought that they had al- ready obtained enough information from Agricola on CD-ROM, and one respon- dent could not find appropriate informa- tion by using Agricola and selected an- other laserdisk database. Twenty-six (58%) of the 45 CD-ROM us- ers who participated were graduate stu- dents, 16 (35%) were undergraduates, and 3 (7%) were university staff or community users. None of the participants was a fac- ulty member. Eighteen (60%) of the 30 us- ers who conducted the online Agricola search were graduate students, 11 (37%) were undergraduates, and 1 (3%) was a university staff or community user. Twenty-seven (90%) of the online updat- 324 College & Research Libraries ing participants indicated that they had used a laserdisk or online database during the previous year. Search Results and Participant Satisfaction The mean number of citations retrieved by the Agricola online updating system was 38, with a median of 1.5. The maxi- mum number of citations retrieved was 395 and 12 ( 40%) of the searches retrieved no citations. Although 19 (63%) of the on- line searches retrieved only 2 or fewer cita- tions, 20 (67%) of the participants thought that they had obtained enough citations to enhance their research efforts. When par- ticipants were asked, ''How would you rate the overall results of online updating of Agricola/CD-ROM?," 25 (83%) gave a rating of ''satisfactory'' or ''very satisfac- tory.'' When asked, ''If online updating for Agricola/CD-ROM was available on a regular basis, how often would you use it?", 24 (80%) participants answered "of- ten" or "always" (see figure 1). Twenty- six participants (87%) indicated that they would prefer ''conducting searches using both Agricola/CD-ROM and the online updating system (with assistance)." Search Costs and Fees Actual search costs were recorded and cost per search ranged from $.42 to $98.97. The mean cost of the online searches was $9.50, with a median cost of $1.82. Twenty-six of the searches (87%) cost less than $10 and 17 (57%) cost less than $2 (see figure 2). Participants were queried about using the Agricola online updating system for a fee, and when asked, ''If online updating for Agricola/CD-ROM was available for a modest fee ($2 to $5), how often would you use it?", 22 (73%) participants an- swered "never," "seldom," or "some- times" (see figure 3). When the same question was rephrased, ''If a fee were charged for online updating how much would you be willing to pay?,'' 100% of the participants indicated that they would pay $5 or less. The responses to both ques- tions indicate that the majority of patrons are only willing to pay under $5 for online updating. July 1990 OBSERVATIONS AND DISCUSSION The information gained from the re- searchers' observations on their subjects' searching behavior was as enlightening as the quantitative data supplied by the questionnaire. Because the researchers monitored the patrons while they did their DIALOG update searches, they were able to get a firsthand look at their search strategies. The observations made con- cerning these searches were revealing and led to the unexpected conclusion that the majority of the participants did not under- stand the basic concepts of searching, such as selection of search terms, use of Boolean operators, truncation, and limit- ing. The participants did not understand the basic concepts of searching, such as selection of search terms, use of Boolean operators, truncation, and limiting. The Evans Library at Texas A&M Uni- versity first offered end-users access to BRS/ After Dark in 1984. Since that time, Knowledge Index and STN have been added. Users are required to read a man- ual before their appointment and to show the attendant their search terms. After dis- cussing their terms and possible search strategy, the patrons are logged on to the appropriate online system. The attendant is available to answer questions at any time during the session. Both printed and personal assistance are available before and during the online search. A similar arrangement exists for the six- teen laserdisk databases available in the Reference Division. Instructional hand- outs prepared by reference librarians or database producers are available at each workstation. Many of the laserdisk sys- tems, such as the Silver Platter databases, have an extensive series of help screens. Most importantly, the area is staffed from 8 a.m. until closing (10 p.m. on weekdays and Sundays) by student assistants and classified staff members. The staff circu- Enhancing CD-ROM Searches 325 Patrone 20---------------------------------, / / 161-................................................................................. . 10 f. ................................................................................ .. / / 6 .......................... p./-~!. ..................................... .. o~~(-~0~~-~~~~-~~-~--£~~ Newr Some Often Always FIGURE 1 How Often Would You Use Online Updating? $10-$100 13.0~ $0-$2 67.0~ FIGURE2 Cost of Searches 326 College & Research Libraries July 1990 Pat rona 16~--------------------------------~ 10 ................................................................................. . NEW8r Seldom Some Often Always FIGURE3 How Often Would You Update for a Small Fee? ? s (land or farm) and leas? 23569 LAND 62832 FARM 1920 LEAS? 51 673 (LAND OR FARM) AND LEAS? ? s stock? 52 4635 STOCK? ? s invest? or £inane? 13887 INVEST? 4499 FINANC? 53 18072 INVEST? OR FINANC? ? s s2 and s3 4635 52 18072 53 54 111 52 and 53 ? s risk? 55 5023 RISK? ? s s 1 or s4 or s5 673 51 111 54 5023 55 56 5773 51 or 54 or 55 ? s s6 and ud=8710:9999 5773 56 65285 UD = 8710 : UD = 9999 57 395 56 AND UD = 8710:9999 FIGURE4 "Farm Investment" Search Strategy late throughout the area providing de- tailed one-on-one instruction. Once again, help is readily available at anytime during the search. Many participants had difficulty with the most basic principle of searching- selection of appropriate search terms. For example, a patron who was looking for ar- ticles on HAPLOIDS IN FORESTRY used the term FOREST as a keyword without applying truncation. Upon further ques- tioning, it became apparent that the use of specific tree names would have been valu- able. Searchers also neglected to use the scientific name (genus and species) of or- ganisms in addition to the common name. Few of the patrons used Boolean opera- tors or if they did, most used them incor- rectly. Out of the 45 searches, fewer than 10 used truncation. Not a single patron limited the search to specific fields (title, descriptor, etc.). The search strategy illus- trated in figure 4 does incorporate the use of truncation and Boolean operators. Un- fortunately, these techniques were used incorrectly (see figure 4). Even when the search was restricted to the latest online updates, 395 citations were found. Despite the availability of onscreen, printed, and personal assistance, many end-users may not be as skillful at search- ing as librarians would like to believe. 5 Some librarians assume that since the pa- trons are not asking for help, they do not need any, and that patrons are finding the material they need in the most efficient manner possible. The observations ob- tained during the course of this study indi- cate that this assumption is not accurate. The search strategy formulation problems encountered during this study are espe- cially significant when taking into consid- eration that 90% of the participants indi- cated on their questionnaires that they had searched a laserdisk or online data- base within the past year. The majority of the participants (60%) were graduate stu- dents with a strong interest in and com- mitment to their research. Yet, it is obvi- ous that these searchers were not doing their online and/or laserdisk searches in the most effective way possible. These end-users, doing both online and ondisk searches, seemed satisfied that they were Enhancing CD-ROM Searches 327 able to find something and find it quickly. They were impressed by the speed of the system and the large number of citations. When using both the online and ondisk systems, some of the participants happily printed out hundreds of citations. These observations also indicate that little atten- tion is given to quality of the search results when the end-user can quickly print the citations and is charged little or nothing for the search. Participants also experienced difficulty when switching between two separate systems such as Silver Platter and DIA- LOG. Logistically it was inconvenient to move from one search station to another. The necessity of using Silver Platter com- mands on the CD-ROM database and then adjusting to DIALOG commands for the online updating also provided partici- pants with a major challenge. Most pa- trons needed a great deal of help. It quickly became apparent that they lacked the theoretical framework to make the transition smoothly. In addition, very few libraries could spare the staff needed to log patrons onto the online database and give them instruction on a second set of commands. This approach would become even more unrealistic if the online updates were provided "on demand" and not by appointment. In our enthusiasm to embrace CD- ROM technology, librarians have ne- glected to make patrons aware of its drawbacks. In our enthusiasm to embrace CD-ROM technology, librarians have neglected to make patrons aware of its drawbacks. Many of the patrons who participated in this research project were surprised to learn that Agricola on CD-ROM was not as current as its online counterpart. This type of response indicates that librarians should alert CD-ROM searchers to the possibility that they are not getting the most current information available. For many graduate students and faculty mem- bers, their ability to obtain the latest mate- 328 College & Research Libraries rial is critical. These patrons especially should be aware of the lack of currency in- herent in CD-ROM databases. One partic- ipant wrote ''As a graduate student, I am encouraged to be on top of all current work in my field. Especially in state-of- the-art, technical research the updated search is a must!'' Libraries must provide patrons access to those recent citations missed by CD-ROM databases by sub- scribing to the online version. The re- sponses provided by the participants of this study support the premise that online databases can be used to complement their CD-ROM counterparts to provide patrons with comprehensive coverage. CONCLUSIONS The majority of participants were eager to update their CD-ROM Agricola search with an online search. They thought that conducting an online search outweighed the time and effort involved. However, in spite of their enthusiasm, patrons were only willing to pay a modest fee ($2 to $5) for the updating service. It is likely that pa- trons would supplement their CD-ROM searches with online searches if the ser- vice were available for a minimal charge. With a majority of searches costing under $5, it may be feasible for libraries to pro- vide online updates to CD-ROM database searches. Patrons probably would use an online updating system only if it were conve- nient and easy to learn. Several currently available systems, such as those produced by Wilson and DIALOG, provide a smooth transition between the CD-ROM July 1990 and online databases. Another advantage is that the same searching commands are used on both types of databases. How- ever, subscribing to a system with auto- matic logon capability or a single com- mand language is not the whole answer. Patrons still need instruction from library staff. As librarians gain more experience with end-users of both online and ondisk databases, it is becoming apparent that these CD-ROM systems are not as self- service as they have been proclaimed to be. Participants in this study, most of whom had searched a database before, and all of whom had access to onscreen, printed, and personal assistance, were not using the system to its fullest capacity. The quantity of citations, speed, low cost, and hard copy seems to overshadow us- ers' concerns for quality of citations. Online databases are necessary comple- ments to laserdisk databases in order to provide patrons with thorough and com- prehensive information. These hybrid search systems will only be practical if they are convenient for patrons to use, available at little or no cost, and utilize the same search command structure. Prod- ucts such as the DIALOG or Wilson sys- tems are representative of the first step that the information industry must make to provide online updating for CD-ROM databases. Additionally, efforts to coordi- nate the use of these resources will be more successful if information profession- als and industry members cooperate to as- sume the responsibility of integrating these services and providing effective in- struction. REFERENCES 1. Rick Meyer, "Strategies for Libraries," Bulletin of the American Society for Information Science 14:22 (Oct./Nov. 1987). 2. Ibid. 3. Vicki Anders, Wiley Laser Disk Seroice: Statistical Report, April1988 (College Station, Tex.: Reference Division, Sterling C. Evans Library, Texas A&M University,© 1988). 4. Vicki Anders, Automated Information Retrieval Seroice: Annual Report 1987/88 (College Station, Tex.: Reference Division, Sterling C. Evans Library, Texas A&M University, © 1988). 5. Stephen P. Harter and Susan M. Jackson, "Optical Disk Systems in Libraries: Problems and Is- sues," RQ 27:519 (Summer 1988).