Journal-citation-seeking behavior at two health sciences libraries

INTRODUCTION

Understanding how patrons in the health sciences identify citations and other information is integral to the mission of health sciences libraries. Studies relying on surveys, focus groups, and interviews reveal only how patrons state they seek information and may not reflect actual behavior. Understanding journal-citation-seeking behavior may help improve tools to deliver health information. This study used direct observation to see how patrons presented article citations to their libraries. Library staff compared the resources patrons used to identify specific journal articles at an academic health sciences library and at a state health department. Observing the general patterns of how patrons present citation requests reveals some of the methods clients use to find information in journals.

LITERATURE REVIEW

Researchers frequently conduct surveys and focus groups to investigate information-seeking behavior of people in the health sciences but rarely observe them as they actually seek information. Responses to questionnaires about information-seeking habits may not reflect actual behavior [1, 2]. Information seekers in the health sciences rely heavily on colleagues and meetings as resources for information according to several surveys [3–8]. Health sciences professionals also frequently browse journals when seeking current information [9, 10]. Life science students report infrequent use of indexes and chaining through bibliographies from other publications when searching for information [11]. Other studies report relatively infrequent use of indexes [12, 13]. If an index is used to locate information, MEDLINE remains the database of choice in the health sciences [14–16]. Focus group studies yield similar results [17–19]. One recent study states that the increased availability of electronic journals has not yet significantly altered the research process [20].

Few observational studies are available regarding the journal-citation-seeking habits of library patrons in the health sciences. Many journal-use studies reflect which materials patrons use in libraries [21–24], and other studies highlight journal use through interlibrary loan or document requests [25, 26]. One study states that library collection size does not correlate to the number of different journals cited by faculty in their work [27]. These studies typically characterize frequently used library materials, but none of them reveal much about how patrons identify specific journal articles.

SETTING AND METHODS

Library staff observed citations that patrons brought to the reference desk at the Bio-Medical Library (Bio-Med) at the University of Minnesota, Twin Cities (UMN) to help understand how patrons identify journal articles for their information needs. Bio-Med is a library in a large urban academic health center with a busy reference desk. The library serves a wide variety of faculty, staff, and students as well as the public. Two reference staff members examined 200 sets of citations brought to the reference desk as patrons asked questions from May to August of 2001. A variety of reference shifts were sampled between the reference hours of 8 a.m. to 7 p.m.

Lists of citations were also examined at the RN Barr Library at the Minnesota Department of Health (MDH) to observe any differences in citation-seeking behavior in a nonacademic setting. RN Barr serves approximately 1,300 public health professionals and support staff throughout the State of Minnesota. The library is located a few blocks from Bio-Med and has a small branch located in downtown St. Paul. Since RN Barr does not have a reference desk, citation sources were recorded as department staff asked questions or dropped off article requests. Library staff examined 200 sets of citations from March to September of 2002. Most in-person requests were used for this study.

Library staff classified how patrons presented a total of 400 requests at the two different health sciences libraries. The origin of each citation was determined and assigned to a category (Table 1). Sets of citations printed from indexes, printed from Websites, or marked on bibliographies from other materials pinpointed original sources. Other groups of citations presented as typed lists, handwritten requests, from email, or from memory often did not indicate the exact original source. The date, citation-origin category, and index name (when applicable) was recorded in a Microsoft Excel spreadsheet.

Table 1 Citation-origin Categories

RESULTS

The ways in which patrons presented their article citation requests was similar in both libraries for most of the categories (Figure 1). Patrons used citation lists printed from indexes 37% of the time. MEDLINE was the index of choice at both locations, accounting for 86.5% of the index citations at RN Barr and 80% at Bio-Med. Patrons at RN Barr presented their citations as typed lists 11% of the time and at Bio-Med 12% of the time. Typed lists included citations generated from bibliographic management software or created in word-processing programs. Patrons found new citations through email correspondence or email discussion lists about 5% to 6% of the time. Very few people relied on their memory at either library to convey their specific citation requests (0%–1%).

Figure 1.

Citation origin

Library staff noted some slight differences among some of the other citation-origin categories. Patrons brought handwritten citation requests on scratch paper to Bio-Med more frequently than to RN Barr (28.5%, 13%). RN Barr patrons relied on chaining through bibliographies in publications they already had to identify additional resources more often than those visiting Bio-Med (24.5%, 14%). Patrons requested articles that they found cited on Websites more frequently at RN Barr than patrons at Bio-Med (8.5%, 0.5%).

DISCUSSION

Index use

The total number of sets of citations printed directly from indexes was the same at Bio-Med and at RN Barr (37%). RN Barr does not provide access to any indexes for their clients outside of free indexes, such as PubMed. Bio-Med supplies a wide variety of databases for patrons and yet exactly the same number of citations was seen printed directly from indexes. According to this data, proportions of patrons using indexes to identify journal articles remains somewhat constant regardless of the number of different indexes available.

The number of different indexes used for seeking journal citations was nearly the same at both libraries. Only 15% to 20% of the index citations viewed at either library originated from indexes outside of MEDLINE. MDH staff, who have extremely limited access to indexes, brought in citations from ten different indexes. Bio-Med offers access to many indexes, yet citations were seen from only one additional index. Access to a wide variety of health sciences indexes for patrons does not seem to guarantee widespread use.

Personal preference, convenience, and availability play a role in MEDLINE interface choice. As expected, RN Barr patrons used PubMed 92% (59 of 64) of the time for access to MEDLINE, as it is the only interface available to them. However, even when OVID MEDLINE was available at Bio-Med, 50.8% (30 of 59) of the MEDLINE citations seen at the reference desk still originated from PubMed. Convenience, access, familiarity, and personal preference clearly affect choice of one MEDLINE product over another [28].

Handwritten citations

Staff observed more handwritten citations at Bio-Med than at RN Barr (28.5%, 13%). The origin of these citations remains unknown. The fact that most MDH employees do most of their searching in their own workspace and have access to printers could account for fewer handwritten citations at RN Barr. In contrast, public access computers at Bio-Med are heavily used. Patrons may find it more convenient to jot something down while they are searching at the library since they must pay and wait for printouts. The number of handwritten citations could indicate additional index use at Bio-Med, but it is impossible to determine the exact origin of these citations.

Website use

RN Barr patrons relied on non-index Websites as a resource for citations more often than those at Bio-Med (9%, 0.5%). These sources included government, nonprofit, and journal Websites. MDH staff may turn more often to the Internet because other library resources are scarce. The fact that a greater proportion of public health staff are practitioners rather than researchers may also contribute to some of this difference. Perhaps bibliographic instruction and the library Website in a university setting steers students towards indexes and other library sources for citations. However, index use was no higher at UMN in this study.

Chaining through bibliographies

Selecting references from a bibliography at the end of a publication was a common practice among those who used either library. One study suggests that 95% of scientists in the physical sciences use this method for locating citations [29]. This research method is also used commonly in the humanities and occasionally among undergraduates [30, 31]. RN Barr patrons brought citations to the library marked on other publication bibliographies 25% of the time, while Bio-Med patrons did this 14% of the time. With limited resources available at RN Barr, perhaps patrons refer frequently to materials in their possession when identifying additional resources to request from the library.

Study limitations

This simple study has some limitations for completely characterizing the journal-citation-seeking habits of patrons in the health sciences. The sample size of 400 total sets of citations is limited, particularly when divided among seven categories at two libraries. Staff viewed 200 sets of citations at each location, but a particular patron may have been surveyed more than once at either location. The sample size was dictated by the time available since each of the two staff members observed citations at both locations. Furthermore, library staff only studied the information-seeking patterns of those who use libraries, and not those who sought information elsewhere.

This study is also limited in that it only measures how citations are presented to libraries as reference questions or as article requests, but not necessarily how patrons found the citations in the first place. The exact origin of many of the sets of citations remains unknown since the categories are not mutually exclusive. For example, the original source of handwritten citations remains unknown; the actual origin could be from any of the other categories. The origins of citations printed from indexes and marked on article bibliographies are more definitive.

CONCLUSION

In spite of the study limitations, some general observations can be made regarding citation-seeking behavior in the health sciences. Few differences were seen between citation-seeking behavior of those at an academic institution and those at a health department. Library patrons still rely on indexes less than half of the time to identify new journal articles. Few indexes besides MEDLINE are used. Patrons obtain citations from publication bibliographies, email correspondence, and from Websites. The size of a library and the availability of library resources do not significantly alter citation-seeking behavior. Patrons in the health sciences probably rely on a combination of several different techniques to identify relevant articles for their work.