College and Research Libraries Student Use of Online Catalogs and Other Information Channels Ingrid Hsieh-Yee Many information channels are available to students, but it is not clear how students select and use them. A survey of juniors at two universi- ties found online catalogs to be the most commonly used information channel. Students identified "convenience" and "quality of data" as the top two factors in their choice of information channel. Not many students accessed the online catalog remotely. Most conducted keyword and subject searches, but few seemed to understand the differences be- tween them. Students searched the online catalog file and the data- bases containing the most recent literature more frequently than other databases, and had positive views about searching the system. Analy- ses of selected characteristics of students showed that race, school, and economic background contributed to their information-seeking be- havior. The study affirms the value of librarians and of formal informa- tion channels, suggests ways to augment the online catalog, and re- minds librarians that equal access to information technology does not necessarily lead to equal use. cademic and research libraries have assimilated information technology into their services to support study and teaching. Besides print tools, students now have online catalogs, electronic databases, CD- ROM databases, and network technology to help them search for information. But it is unclear how students who are under time constraints and academic pressure make use of the information channels available to them. Do they prefer one over the others? What do they take into ac- count in selecting a channel? How do they use the online catalog, which has the most investment from the libraries? And what factors contribute to their use of informa- will enable librarians to plan f?r ~ost-ef­ Tective services allil will suggest new fea- ~ures to make the online catalog more ~ user-friendly. This study attempts to ad- dress These questions. Literature Review Information-Seeking Behavior of College Students Researchers have analyzed the informa- tion-seeking behavior of undergraduate students from various perspectives, have examined the problems of library use, and have categorized library anxiety, while exploring the mental states of students in using library resources. 1•3 Students' use of one or more information sources is often tion channels? Answer th~se qu~s~o~s the focus of investigation, as is their use Ingrid Hsieh-Yee is an A ssistant Professor in the School of Library and Information Science at Th e Catho- lic University of America; e-mail: hsiehyee@cua.edu . 161 162 College & Research Libraries of library collections, which has received more attention from researchers than other topics. Several studies confirm that college students make little use of library collections. 4 Researchers also have investigated possible causes of students' choice of in- formation channels, but the results are less conclusive. Kathleen Dunn reported that undergraduates made considerable use of informal, interpersonal sources such as family members and friends, but ranked teachers first and the library sec- ond in terms of their importance in pro- viding information. She also found that students interested in intellectual stimu- lation and professional success tended to use libraries and experts for needed in- formation.5 Tony Mays reported a lack of a clear correlation between academic achievement and use of library collections in Australia, and concluded that under- graduates tend to regard libraries as study halls and consider them superflu- ous to their educational program.6 Nigel Ford, however, found that some students regarded libraries as an extension of what they learn from lectures.7 Mays also re- ported that no prediction can be made between discipline and library use.8 How- ever, Stephen A. Osiobe found that disci- pline related to Nigerian undergraduates' use of information channels, though the relationship was very weak.9 As for stu- dents' use of information sources, Osiobe found browsing to be most popular, with faculty and the card catalog tying for sec- ond, followed by librarians, references in articles and books, abstracts and indexes, and colleagues. Focusing on students' progress through the research process, Barbara Fister reported that they used online catalogs, indexes, and reference tools to browse the field and considered citation network the most direct and effi- cient way to find relevant information. 10 College Students and Online Catalogs \/Because academic libraries were among the first libraries to make online catalogs March 1996 available, much of the literature on on- line catalogs examines undergraduates' use of online catalogs. College students are reportedly enthusiastic about online catalogs and prefer them to card cata- logs.11·12 They conduct more subject searches on online catalogs but experi- ence difficulties with these searches.B Nor do they have a clear understanding of the contents of an online catalog. 14 The avail- ability of keyword access has made it the preferred searching method. For instance, Pat Ensor found that most users (73%) at Indiana State University use keyword searching. 15 She also reported that many users (50 %), like those interviewed by Karen Markey, misunderstand the nature of subject searches and do not know that they need to use the Library of Congress Subject Headings (LCSH) for subject searching. 16·17 Catalog Improvement and Expansion Researchers have explored various meth- ods to enhance or improve subject searches. Markey and Anh N . Demeyer, for instance, studied the use of the Dewey Decimal classification scheme for re- trieval.18 Jeffrey C. Heustis analyzed LC classification numbers for browsing and Marcia J. Bates suggested incorporating a "superthesaurus" into online cata- logs.1 9·20 Mary Micco developed a proto- type catalog based on subject clusters while Diane Vizine-Goetz and Karen Markey Drabenstott examined whether matching subject terms with LCSH was the best way to retrieve items relevant to the user's topic.21·22 Ray R. Larson experi- mented with classification clustering in CHESHIRE. 23 In the meantime, several online catalogs expanded to provide us- ers with access to collections of other li- braries and commercially produced da- tabases. However, it is unclear whether the use of such augmented catalogs is the same as that of the earlier online catalogs. Research Questions The present study investigated the infor- mation-seeking behavior of students at the American University (AU) and the University of the District of Columbia (UDC) because their student bodies offer an opportunity to examine the relation- ships between information-seeking be- havior and factors such as race and eco- nomic background. Specifically, the re- search questions were: • What are the information channels used by college students when they work on course-related projects? What factors do students consider in selecting infor- mation channels? • How do college students use ALADIN, an augmented catalog that pro- vides access to the holdings of the Wash- ington Research Library Consortium and several commercial databases? • How are selected characteristics of students related to their use of informa- tion channels, ALADIN and libraries? Methodology Population and Sample The study's population included juniors at AU and UDC because these students could be assumed to have had some ex- perience in looking for information after completing two years of college educa- tion. By estimating that 75 percent of the subjects will display different informa- tion-seeking behaviors and setting the confidence level in the findings at 95 per- cent, the author, using a statistical pro- gram (StatPac Gold), found the desirable sample size to be 256. A random sample of 11 percent of the population was then taken, resulting in a sample of 258 jun- iors. Survey Instrument Because information on college students' information channels is sketchy, the li- brarians who assisted with the project held two focus group interviews to iden- tify all the channels students explored when they worked on an assignment. Stu- dents described how they selected an in- formation channel, used ALADIN and Student Use of Online Catalogs 163 libraries, and offered suggestions for im- proving library services. The author de- veloped a questionnaire from these find- ings, pretested it on twenty-four students in the study population, solicited com- ments from the librarians, and revised it for the main study. (The survey instru- ment is available from the author upon request.) Data Collection The original plan involved collecting data through a mail survey. The author guar- anteed subjects anonymity and provided a self-addressed, stamped return enve- lope for their responses. Because of a de- lay in reviewing the study at one univer- sity, the first main test did not take place until mid-April of 1994. By then, final examinations were approaching and stu- dents did not respond well to the survey. The author then obtained an extension of the deadline from the funding agency so that the research team could collect bet- ter data. During the summer, the author revised the survey instrument again and the team decided to collect survey re- sponses on campus in addition to the mail survey. The librarians agreed to post signs around their libraries to encourage par- ticipation or to seek faculty assistance in collecting data. Researchers mailed 258 surveys in Sep- tember of 1994 and mailed another set of surveys in October to increase the re- TABLEt Survey Response Rates UDC AU Row Total Sent 137 121 258 Received 83 74 157 Response rate 61% 61% 61% Undelivered mail 34 9 43 Usable responses 49 65 114 Response rate 48% 58% 53% (Undelivered ones removed) 164 College & Research Libraries TABLE2 Use of Information Channels Channel Percentage of Responses• Online catalogs 88 Roommates/classmates/friends 71 References (books or articles) 70 Librarians 70 Printed indexes 63 Stack browsing 61 Reserved readings 61 Teachers 57 Agencies 50 Students who took the course 46 Special bibliography 46 Parents 42 CD-ROM databases 28 Online databases 19 •Respondents reported all the information channels they have used, so the total for each channel is potentially 100%. sponse rate. By mid-November, research- ers received a total of 157 responses, re- sulting in a response rate of 61 percent. After excluding incomplete and unde- livered surveys, the response rate went down to 53 percent, with a total of 114 usable survey (see table 1). Data collection on campus met with varying success. AU contributed forty- TABLE3 Frequently Used Information Channels Most frequently used channels Percentage Online catalogs 53 Print indexes 11 Teachers 11 Second most frequently used channels Online catalogs 16 References (books or articles) 14 Print indexes 12 March 1996 four completed surveys, and UDC twenty- five surveys. To determine whether cam- pus respondents and mail respondents displayed different behaviors, the author performed Chi-square tests on forty-four selected dependent variables. Test results show that the two groups differed in only two variables, so the researchers con- cluded that no separate analyses were necessary. The final analysis included sev- enty-four surveys from UDC and 109 sur- veys from AU. Data Analysis A research assistant coded data and en- tered them into StatPac Gold, and the au- thor compiled descriptive statistics and performed Chi-square tests to analyze the associations between variables. Findings Profiles of Respondents Forty percent of the respondents were enrolled in UDC, and the rest in AU. Nearly 60 percent were female, and 39 percent male. Whites represented 51 per- cent of the respondents, African Ameri- cans 29 percent, and Asians and Hispan- ics ten percent. The other ten percent did not specify their race. As for economic background, slightly more than 52 per- cent of the respondents came from fami- lies with an annual income of more than $35,000, and 65 percent of the respondents had a part-time job. Very few of the re- spondents (18%) had children. Findings on Information Sources Information channels. Students identi- fied fourteen information channels dur- ing the focus group ~terviews. Most sur- vey respondents (88%) said they used online catalogs, and at least 63 percent reported consulting roommates and friends, book and article references, librar- ians, and printed indexes. It is somewhat surprising that only 19 percent of the stu- dents used online databases and 28 per- cent used CD-ROMs. Such low use, fur- thermore, was not because of students' Student Use of Online Catalogs 165 TABLE4 Factors Affecting Information Channel Selection Factor Convenience The source is easy for me to get to Quality of data I generally find very good information this way Ease of use Availability Experience I used this method many times in the past Cost Score• 3.43 4.02 4.66 4.80 5.70 7.08 Ranking 1 2 3 4 5 6 "The score is the mean of values assigned to each factor. The lower the score, the more important the factor is to respondents. lack of knowledge-more than half of the respondents had knowledge of CD-ROM databases, 43 percent had searched them, and 44 percent knew about online data- bases. In spite of the rapid growth of on- line databases and CD-ROM titles, few students used them for needed informa- tion. Table 2 presents students' use of in- formation channels. As table 3 shows, among the 14 chan- nels, students showed a clear preference for online catalogs. More than half of the respondents listed them as the most fre- quently used method, and 16 percent re- ported them the second most frequently used method. Students also consulted print indexes, teachers or citation refer- ences, but much less frequently than on- line catalogs. Factors in channel selection. Of the six factors that may have affected their se- lection of information channels, students identified "convenience" and "quality of data" as the top two. "Ease of use" ranked third, followed by "availability," "expe- rience," and "cost" (see table 4). Experience with ALADIN Learning ALADIN. Students learned to use ALADIN through various methods. were less likely to read the printed instructions or seek help from a friend. The strong reliance on on-screen instructions underscores the importance of system interface, screen displays, and message design. As the system evolves, improvement in these areas will encour- age its use. Remote access. Considering the wide- spread use of personal computers, it is somewhat surprising that only 28 percent of the respondents accessed ALADIN from outside the library. But this finding is consistent with earlier reports that most undergraduates were in-library users and that faculty, graduate students, and other academic staff tended to be remote us- ers.24·25 The reason for the small number of remote student users could be that most students did their research in libraries. Or, it could be that students could not afford a modern or might not know how to use one. As the Internet becomes more popu- lar among students, it will be interesting to monitor whether remote access to ALADIN increases. Useful bibliographic elements. Stu- dents reported three elements that they usually looked for on an ALADIN dis- play. They presented a variety of patterns, with "author-title-subject" being the most 166 College & Research Libraries TABLES Sources of Keywords and Subject Headings Percentagea Source of keywords What I can think of at the time 78 Earlier readings 57 Records I have retrieved 49 Librarian's help 26 Source of subject headings What I can think of at the time 66 Records I have retrieved 57 Earlier readings 54 Librarian's help 21 LCSH 15 3 Multiple sources could be reported, so each source is potentially 100%. common combination (reported by 20%), followed by "subject-call number-status of the book" (9%). Data show that "sub- ject headings" were the most consulted elements, followed by titles, authors, sta- tus of the book(s), and call numbers. Stu- dents could be using the subject headings to determine the relevance of the records or to expand their search. Currently, ALADIN's default display presents no subject headings. The addition of subject headings to the screen will make the sys- tem more helpful to users. ALADIN search method. Of the five search methods, respondents identified "keyword" as the most frequent method for access. "Subject headings" ranked sec- ond, and "author" and "title" tied for the third. A "call number" search was the least used. Consistent with earlier reports that users relied more on keyword search- ing and tried more subject searches in the online environment, survey respondents searched by keyword and subject head- ing more often than author, title, or clas- sification number. Keywords versus subject headings. Findings confirm librarians' concern that users may not understand the differences March 1996 between a keyword search and a subject heading search. Although 96 percent of the respondents said they understood the differences, only 15 percent said they used LCSH to find subject headings and 66 percent said, "I use whatever I can think of at the time." The reported sources of their keywords and subject headings re- veal a similar pattern and suggest that many students probably did not know the differences between keyword searches and subject heading searches (see table 5). Used properly, subject headings en- able a searcher to retrieve relevant items quickly, but it is difficult to come up with a subject heading without consulting an authority list, which for most online cata- logs is LCSH. It is therefore surprising that few students sought librarian's as- sistance in finding subject headings and even fewer consulted LCSH. One possible explanation is that students did not know that they needed to use proper subject headings in subject heading searches. One encouraging sign, however, is that 57 percent of the students turned to the retrieved records for subject heading ideas, although it is unclear whether they examined the subject headings area or the entire records for that purpose. Because students displayed a strong interest in topical searches and many of them con- ducted subject heading searches with keywords, improvement in this area is critical. As a start, the system could pro- vide on-screen explanations and ex- amples that illustrate the differences be- tween keyword searching and subject heading searching, and librarians could highlight such differences at training workshops. At a more sophisticated level, the system could provide browsing of authorized terms and cross-references, and allow users to find items similar to the records they found. 26 Advanced search features. Earlier studies reported that advanced features . on online catalogs were seldom used. Re- spondents of this study did use advanced Student Use of Online Catalogs 167 TABLE6 features, but the use was low. Fifty-six percent of the students used Boolean op- erators and 37 percent used limiters, but only 21 percent used truncation. It is en- couraging that more than half of the respondents used Boolean operators for their searches, suggesting that students are becom- ing more sophisticated in searching. Librarians, none- theless, should continue to educate users of the value of Boolean operators be- cause these operators can increase the relevance of re- Views on ALADIN Searches by Percent View Agree(%) Disagree(%) Title search is easy 67 7 Author search is easy 66 11 Subject heading search is easy 48 19 Keyword search is easy 60 14 Call number search is easy• 20 19 Moving back and forth is easy 69 13 Expanding a search is easyb 31 30 Reducing a search is easyc 42 22 Note: Four responses were provided, including "agree," "neutral," "disagree," and "don't know," but only "agree" and "disagree" are reported here. "41% reported not knowing whether call number search is easy. bll% reported not knowing whether expanding a search is easy. ell% reported not knowing whether reducing a search is easy. trieved records, if used properly. Because keyword searches tend to retrieve a large number of records, it is understandable that trun- cation was not used often to increase re- trieval. But the need to narrow retrieval is clear. At the focus group interviews, students said they often preferred items owned by their libraries, but only few stu- dents knew how to limit their search by location. Judging from their enthusiasm for this feature, students are likely to ap- preciate some instruction on limiters. Files searched. Of the nine files in ALADIN, the online catalog file was the most frequently used, followed by the multisubject periodical index covering literature from 1990 to the present time. The other files such as Newspaper Ab- stracts, ERIC, Periodical Articles 1983- 1993, and ABI/Inform were used infre- quently. It seems that a catalog to local holdings and an index to recent literature would satisfy most students' needs and that undergraduates had little use for older literature or more specialized sub- ject files. Views on ALADIN searches. The sur- vey presented descriptions of ALADIN searches with different phrasings to keep respondents from falling into a pattern in responding to them. To make the statis- tics easier to understand, table 6 rephrases these views in presenting students re- sponses. As table 6 illustrates, students' views on ALADIN searches were generally positive. At least 60 percent of the respon- dents found title, author, and keyword searches easy, whereas more than 40 per- cent found subject heading searches easy and reducing a search result easy. Forty- one percent did not know whether call number searches were easy, and those who attempted to search by call number were split on its difficulty. Students were similarly divided on the difficulty of ex- panding a search when they retrieved too little information the first time. Another way to analyze the data is to focus on the activities with which users experienced difficulty. Search expansion was difficult for one-third of the respon- dents; subject heading search, call num- ber search, and search reduction posed problems for nearly one-fifth of them; and very few students knew how to conduct call number searches. Considering the strong interest in conducting subject searches, librarians will be doing a great service to students by providing more instruction on these types of searches. 168 College & Research Libraries TABLE7 Views on Research Statement Agree(%) Research is critical 96 to my future career Everyone should know 68 about research It's nice to know about it, 12 but I don't have the time to do that I would like to know more about 69 it so that I am well prepared for advanced studies I can always rely on 4 librarians later I want to do the research myself 52 Experience with Libraries Student responses to questions on library experience show that they viewed librar- ies as an important part of their educa- tion. They ranked "to work on a paper or an assignment" first and "to meet friends" and "to rest" last as the reasons for using libraries. Sixty percent of the students thought that they knew how to make good use of library resources, whereas 40 percent responded negatively. Librarians can easily address this problem by pro- viding the kinds of assistance students need, such as "information about library collection and services" (57%), "instruc- tion on how to prepare research papers" (49%), instructions for using the library (40%), and a map of the library (38%). Although 60 percent of the respon- dents knew how to make good use of li- TABLE 8 March 1996 brary resources, only 46 percent reported success in finding what they needed. The main reasons for the search failure were "holdings failure" (88%) and "items not on the shelf" (69%). Most students displayed confidence in their understanding of research (44% "good understanding," 51% "some un- derstanding"). As table 7 illustrates, a great majority of them understood the importance of research and more than two-thirds wanted to know more about it. Roles of Selected Factors The author analyzed the relationships between students' .characteristics and their information-seeking behavior by performing Chi-square tests on these variables. The independent variables were race, school, economic background, gender, part-time job, and children. The forty-four dependent variables included the five most popular information chan- nels, three major factors in information channel selection, two most popular methods in learning to use ALADIN, fre- quency of ALA DIN use, remote access to ALADIN, understanding of differences between keyword and subject heading searches, sources of keywords for search- ing, sources of subject headings for searching, use of advanced search fea- tures, views on ALADIN searches, abil- ity to make good use of library resources, success in finding needed information in libraries, understanding of research pro- cess, views on research, and reasons for using the library. Use of Information Channel Associated with Race Chi-square tests asso- ciate race with :nineteen dependent variables, school with fifteen, and economic background with nine. Some depen- dent variables were as- sociated with more than one independent var~ able. Ideally, multiple Variable Chi-square Cramer'sV Probability Parents 7.521 0.204 0.023 References 7.912 0.209 0.019 Print indexes 21.861 0.347 0.000 ALADIN 13.872 0.276 0.001 p < .05 L__ _____________________ ___, c orre la tion ana 1 yses Student Use of Online Catalogs 169 TABLE9 Use of ALADIN Associated with Race Variable Chi-square Cramer's V Probability Learning ALADIN use 29.734 0.425 0.000 from library staff Frequency of ALADIN use 17.017 Remote access 14.425 Using whatever I think of 18.528 for keywords Using retrieved records 7.465 to find keywords Using whatever I think 11.690 of for subject headings Use of Boolean operators 14.855 Keyword search is easy 19.567 Moving between screens 14.052 is easy Expanding a search is difficult 14.113 Reducing a search is easy 16.602 p < .05 could be used to determine how much each independent variable contributes to the variability of a dependent vari- able, but because all the data were nomi- nal data, such analyses were not possible. The n~xt section discusses the roles of im- portant factors in detail and summarizes findings on less impqrtant factors. Race As a Factor The study analyzed the race factor by Whites, African Americans, and others (Asians and Hispanics combined). Tables 8 through 10 present nineteen depen- dent variables found to relate to the race factor. Table 8 shows that four of the five most reported information channels relate to the race factor. Data suggest that African American students were more likely than the other two groups to ask their parents for information and to rely on references in books and articles for information. They also were the most likely to use print indexes, whereas White students were the least likely to use them. The use of ALADIN showed a reversed pattern. 0.231 0.002 0.296 0.001 0.336 0.000 0.213 0.024 0.267 0.003 0.309 0.001 0.247 0.003 0.210 0.029 0.211 0.028 0.228 0.011 These findings are somewhat troubling because they suggest that African Ameri- can students have not taken full advan- tage of information technology. Because most of the White students are enrolling in AU and the African American students in UDC, it could be that AU's strong li- brary support contributes to its students' heavy use of ALADIN. Or, it could be that other factors such as time or family may have kept some students from using ALADIN. Test statistics suggest race to be a factor, but further studies are necessary to understand fully its effects on students' use of online catalogs. The three groups showed no difference in their reliance on roommates or friends for ideas. Nor were they different in their views on the factors that affected their choice of information channel; all three groups ranked "convenience," "quality of data," and "ease of use" as the top three factors. Table 9 shows that half of the twenty- two dependent variables on students' ALADIN experience related to the race factor. Available data show that 25 per- 170 College & Research Libraries March 1996 TABLE tO Experience with Library Associated with Race Variable Chi-square Cramer 's V Probability Knowing how to make 9.839 good use of library resources Understanding the 14.035 research process Understanding 8.785 research process is critical to my future career Lack of time in learning 6.128 more about research process p < .05 cent of White students, 53 percent of the Asian-Hispanic group, and 73 percent of African Americans consulted library staff. Clearly, all users need some kind of as- sistance to make good use of the online catalog, and the system is not as user- friendly as librarians had hoped. In their use of ALADIN, students also showed different patterns. White students were most likely to: • use ALADIN every time they vis- ited a library, • access ALADIN remotely, • use whatever they can think of at the time for keyword searches, • examine retrieved records for key- words, • use whatever they can think of at the time for subject heading search, • use Boolean operators in their ALADIN searches, • consider keyword searching easy, • · consider moving between screens easy, • have little difficulty in expanding searches, · • consider reducing a search result easy. But African American students were least likely to use ALADIN in these ways. Race, however, was not the sole contribu- 0.238 0.007 0.200 0.007 0.232 0.012 0.194 0.047 tor to the different behavior in using ALADIN, because other factors such as economic background also are involved. Table 10 shows that only four of the fourteen dependent variables on stu- dents' experience with their libraries re- lated to the race factor. Data show that fewer students of the Asian-Hispanic group thought they knew how to make good use of library resources and that more students of this group reported little understanding of the research process than the other two groups. The three groups differed little in their views on research. They agreed that re- search is critical to their future career, but more African American students and stu- dents of the Asian-Hispanic group said they did not have time to learn about re- search. The three groups thought every- one should know about research, ex- pressed a desire to conduct their own 're- search, and shared an understanding that knowing how to do research is important for advanced studies. School As a Factor Table 11 shows that fifteen of the forty- four dependent variables related to the school variable. Perhaps because most White students were from AU and most Student Use of Online Catalogs 171 TABLE 11 Dependent Variables Associated with School Variable Chi-square Phi orva Probability Use of Information Channels References 8.65 0.218 0.003 Print indexes 24.226 0.365 0.000 ALADIN 11.646 0.253 0.001 Exnerience with ALADIN Learning ALADIN from 17.548 0.326 0.000 library staff Frequency of library use 9.426 0.243 0.009 Remote access to ALADIN 22.79 0.372 0.000 Using whatever I can think 13.351 0.285 0.000 of for keyword searches Using whatever I can think 14.802 0.3 0.000 of for subject heading Using LCSH 5.116 0.177 0.024 Using Boolean operators 9.25 0.244 0.002 Keyword search is easy 9.38 0.242 (V) 0.025 Moving between screens is easy 9.761 0.247 (V) 0.021 Expanding a search is hard 10.498 0.257 (V) 0.015 Exnerience with Libraries Understanding research 11.939 0.261 (V) 0.003 process Interest in learning 5.564 0.185 (P) 0.018 more about research 3 Phi and Cramer's V measure the strength of the relationship between two variables. Phi is calculated for two-by-two tables, and V is calculated for tables larger than two-by-two. The value for Vis marked by (V). p< .05 African American students were from UDC, some findings on the school vari- able are similar to those on the race vari- able. Of the five most popular informa- tion channels, test statistics show that UDC students were more likely to use references and print indexes and AU stu- dents were more likely to use ALADIN. Regarding the use of ALADIN, stu- dents displayed the patterns reported under the race variable. The two groups also differed in how they viewed the re- search process. Data suggest that AU stu- dents had more confidence in their un- derstanding of the research process and UDC students were more interested in learning more about research. Economic Background As a Factor The author coded respondents into three categories to analyze the economic back- ground factor: from families with an in- come of less than $30,000, those with an income of more than $30,000, and those with no income information. As table 12 illustrates, the three groups showed no difference in their use of information channels or their assessment of the fac- tors that affected their selection of infor- mation channels. They showed more dif- 172 College & Research Libraries March 1996 TABLE12 Dependent Variables Associated with Economic Background Variable Chi-square Cramer'sV Probability Learning from on-screen 7.969 0.22 0.019 instructions Frequency of ALADIN use 13.927 0.209 0.008 Remote access to ALADIN 6.535 0.20 0.038 Using retrieved records 8.002 0.222 0.018 for keywords Using LCSH 9.204 0.238 0.01 Using truncation 6.999 0.22 0.03 Reducing a search is easy 22.046 0.263 0.001 Understanding of 10.195 0.171 0.037 research process Doing one's own research 10.734 0.257 0.005 p < .05 ferences in their experience with ALA DIN and their experience with libraries. Test statistics suggest that the most affluent group was most likely to use ALADIN every time it visited a library, the less af- fluent group was more likely to use ALADIN sometimes, and the no-response group used ALADIN only when an as- signment needed to be done. Except for access from campus computers, remote access requires some financial commit- ment, and it is therefore not surprising that the groups differed in this activity. As for their views of research, the no-response group contributed most to the differences by having more members reporting "little understanding" and showing no interest in doing their own research. Summary of Findings on Gender, Part-Time Job, and Children Nearly 60 percent of the respondents were female. Test results show that they were more likely to consult parents than male students were and that, in searching ALADIN, male students were more likely to use Boolean operators. Sixty-seven per- cent of the respondents reported having a part-time job. Test results indicate that these students were more likely to use ALA DIN every time they visited a library and to use the library only when a paper was due. They had less success, however, in searching ALADIN than others. Only 18 percent of the respondents had one to five children, and these respondents were more likely to use print indexes and seek help from library staff in learning ALADIN than those without children. Variables Affected by Multiple Independent Variables Seventeen dependent variables related to more than one independent variable. As table 13 shows, race was an important factor that related to most of the seven- teen dependent variables. It also ap- peared often with school or economic background to contribute to students' use of information channels, experience with ALADIN, and experience with libraries. Implications Validation of Established lnfonnation Channels At least 70 percent of the students used online catalogs, book and article refer- ences, and librarians, affirming the im- portance of these channels. Because stu- dents favored channels that were easy for Student Use of Online Catalogs 173 TABLE13 Dependent Variables Associated with Multiple Factors Dependent Variable Variable on Information Channels Parents References Print indexes Online catalogs Variable on Use of ALADIN Learning ALADIN from library staff Frequency of ALADIN use Remote access to ALADIN Using whatever I can think of for keywords Using retrieved records for keywords Using LCSH Using whatever I can think of for subject headings Using Boolean operators Keyword search is easy Moving between screens is easy Expanding a search is difficult Reducing a search is easy Variable on Views of Research Understanding of research process them to get to and that contain good in- formation, academic libraries should con- tinue to invest in online catalogs by aug- menting their contents with high-quality data. Research and academic institutions also should invest in librarians who have effectively bridged users to information sources. Librarians' work is much appre- ciated by college students, and with the incorporation of information technologies into higher education, their roles will probably become more important. Improvements to Augmented Catalog, Such As ALADIN Because many students preferred explor- ing ALADIN by themselves, librarians should design online instruction that fa- cilitates self-teaching. The default screen display should include subject headings so that more users will be able to take advantage of the subject analysis librar- Factor Race, Gender, Children Race, School Race, School, Children Race, School Race, School, Children Race, School, Economic Background, Part-Time Job Race, School, Economic Background Race, School Race, Economic Background, Gender Race, Economic Background Race, School School, Gender Race, School Race, School Race, School Race, Economic Background Race, School, Economic Background ians provide. In addition, librarians need to develop instruction sheets or an online demonstration module that will illus- trate the differences between keyword searches and subject heading searches, and will enable users to decide when a particular type of search will be most ef- fective. To make search expansion and re- duction easier, more context-specific help messages should be provided. As for clas- sification number searches, because most users consider call numbers the addresses of items in their collection and few con- sider them access points, librarians can help them appreciate the retrieval power of classification numbers through instruc- tion workshops or online instruction. Enhanced Library Services and Instruc- tion on Research Data indicate that students use libraries for serious purposes. Most of them knew 17 4 College & Research Libraries how to make good use of library re- sources and indicated that collection fail- ure was the main reason for search fail- ure. Librarians can address this issue by improving document delivery and stack management. Because most students ap- preciate the importance of research and want to conduct their own, librarians can develop courses on information manage- ment to introduce students to the research process, the variety of relevant informa- tion tools, and the skills to synthesize in- formation from various sources. Use of Information Technology and Race Librarians realize that technology could widen the gap between information haves and information have-nots, and, therefore, advocate equal access for all us- ers. Because this study found that more affluent students tend to use ALADIN more often, less affluent students would certainly benefit from librarians' advo- cacy for equal access. But the study also found that equal access does not lead to equal use. Despite the same access to the same information system, White students used ALADIN the most and African American students the least. Test results indicate that many factors contributed to the low use of ALADIN, and there are probably many more factors related to this phenomenon. The study identified factors related to students' use of infor- mation channels. The exact contribution March 1996 of various factors and their interaction await further investigation. Conclusion Although the study focused on two uni- versities in a particular city, several find- ings are consistent with earlier studies and confirm librarians' opinions and views about student users. In addition, the study sheds light on the range of in- formation channels college students use, their selection criteria, their use of an on- line catalog, their self-assessment of li- brary knowledge and research, and the relationships between six f~ctors and their use of information channels. The re- searchers hope that these findings will en- courage library administrators to develop programs and services that will facilitate undergraduate students' information- seeking, provide system designers and managers with ideas for improving infor- mation systems, and prompt administra- tors of higher education to address the special needs of various user groups. Note: This study was supported by a Co- operative Grant for Faculty from the Consor- tium of Universities of the Washington Met- ropolitan Area. The project was conducted with the assistance of Albertine C. Johnson, Chairperson of the Public Services Depart- ment at the University of the District of Co- lumbia, and Elizabeth B. Nibley, Reference Li- brarian at the American University. Notes 1. See, for example, Larry Hardesty, Student Use of the Libraries at DePauw University, 1980, ERIC, ED 187335; and Tony Mays, "Do Undergraduates Need Their Libraries?" Australian Aca- demic and Research Libraries 17 (June 1986): 51-62. 2. Judith Andrews, "An Exploration of Students' Library Use Problems," Library Review 40 (1991): 5-14. 3. Constance Mellon, "Library Anxiety: A Grounded Theory and Its Development," College & Research Libraries 47 (Mar. 1991): 160-65. 4. Nigel Ford, "Psychological Determinants of Information Needs: A Small-Scale Study of Higher Education Students," Journal of Librarianship 18 (Jan. 1986): 47-62. 5. Kathleen Dunn, "Psychological Needs and Source Linkages in Undergraduate Informa- tion-Seeking Behavior," College & Research Libraries 47 (Sept. 1986): 475-81. 6. Mays, "Do Undergraduates Need Their Libraries?" passim. 7. Ford, "Psychological Determinants," passim. 8. Mays, "Do Undergraduates Need Their Libraries?" passim. Student Use of Online Catalogs 175 9. Stephen A. Osiobe, "Information Seeking Behavior," International Library Review 20 (July 1988): 337-46. 10. Barbara Fister, "The Research Processes of Undergraduate Students," Journal of Academic Librarianship 18 (July 1992): 163-69. 11. See, for example, Kenneth Dowlin, "Online Catalog User Acceptance Survey," RQ 20 (fall 1980): 44-47; Carole Weiss Moore, "User Reactions to Online Catalogs: An Exploratory Study," College & Research Libraries 42 (July 1981): 29.:?-302; Joseph R. Matthews, Gary S. Lawrence, Douglas K. Ferguson, eds., Using Online Catalogs: A Nationwide Survey (New York: Neal-Schuman, 1983); and David Steinberg and Paul Metz, "User Response to and Knowledge about an Online Catalog," College & Research Libraries 46 (Jan. 1984): 66-70. 12. See, for example, Sue Pease and Mary Noel Gouke, "Patterns of Use in an Online Catalog and a Card Catalog," College & Research Libraries 43 (July 1982): 279-91; Carolyn 0. Frost, Stu- dent and Faculty Subject Searching in a University Online Public Catalog, 1985, ERIC, ED 264872; Micheline Hancock, "Subject Searching Behavior at the Library Catalogue and at the Shelves: Implications for Online Interactive Catalogues," Journal of Documentation 43 (Dec. 1987): 303-21. 13. See, for example, Karen Markey, "Thus Spake the OPAC User," Information Technology and Libraries 2 (Dec. 1983): 381-87; David W. Lewis, "Research on the Use of Online Catalogs and Its Implications for Library Practice," Journal of Academic Librarianship 13 (July 1987): 152-57. 14. See, for example, Sammy R. Alzofon and Noelle Van Pulis, ''Patterns of Searching and Success Rates in an Online Public Access Catalog," College & Research Libraries 45 (May 1984):110- 15; Matthews, Lawrence, and Ferguson, Using Online Catalogs; and Steinberg and Metz, "User Response." 15. Pat Ensor, "User Characteristics of Keyword Searching in an OPAC," College & Research Libraries 53 (Jan. 1992): 72-80. 16. Karen Markey, Online Catalog Use: Results of Surveys and Focus Group Interviews in Several Libraries (Dublin, Ohio: OCLC, 1983). 17. Pat Ensor, Keyword/Boolean Searching on an Online Public Access Catalog: Patrons and Their Perceptions, 1990, ERIC, ED 337163. 18. Karen Markey and Anh N. Demeyer, Dewey Decimal Classification Online Project: Evalua- tion of a Library Schedule and Index Integrated into the Subject Searching Capabilities of an Online Catalog (Dublin, Ohio: OCLC, 1986). 19. Jeffrey C. Heustis, "Clustering LC Classification Numbers in an Online Catalog for Im- proved Browsability," Information Technology and Libraries 7 (Dec. 1988): 381-93. 20. Marcia J. Bates, "Rethinking Subject Cataloging in the Online Environment," Library Re- sources & Technical Services 33 (Oct. 1989): 400-12. 21. Mary Micco and Rich Popp, "Improving Library Subject Access (ILSA): A Theory of Clus- tering Based in Classification," Library Hi Tech 12 (1994): 5~6. 22. Diane Vizine-Goetz and Karen Markey Drabenstott, "Computer and Manual Analysis of Subject Terms Entered by Online Catalog Users," in Proceedings of the 54th ASIS Annual Meeting -(Medford, N.J.: Learned Information, 1991), 156-61. 23. Ray R. Larson, "Classification Clustering, Probabilistic Information Retrieval, and the Online Catalog, "Library Quarterly 61 (Apr. 1991): 133-73. 24. Evelyn Margaret Kirensen, "Studying Dial-Up Use at UC Berkeley," DLA Bulletin 7 (Dec. 1987): 12-14. 25. Pamela Snelson, "Remote Users of OPACs: Do They Differ from Library Users?" in Pro- ceedings of the 1994 Conference on Integrated Online Library Systems (Medford, N.J.: Learned Infor- mation, 1994), 179-83. 26. INNOPAC and NLS, an in-house system developed for the University of Wisconsin-Madi- son, both provide searchers the capability to retrieve more items directly from subject headings displayed on the screen. For subject searches, INNOPAC also displays authorized subject head- ings and cross-references. Association of and Research College Libraries Why did nearly 11,000 acade:mic librarians and staff join ACRL? World-class publications Standards, Guidelines Advisory services Leadership roles Networking Discounts Sections Advocacy Newsletters Local chapters Career resources Awards, Fellowships Professional development Electronic communications Find out what ACRL can do for you! Call: 1-800-545-2433, ext. 2521 or 2510 E-mail: ctaylor@ala.org or nhale@ala.org