Previous   Contents   Next
Issues in Science and Technology Librarianship
Spring 2002


Making Choices: Factors in the Selection of Information Resources Among Science Faculty at the University of Michigan
Results of a Survey Conducted July-September, 2000

Jane Quigley
Physical Sciences Librarian
Dartmouth College

David R. Peck
Librarian for Chemistry
University of Michigan

Sara Rutter
Librarian for Biology, Natural Resources, and Astronomy
University of Michigan

Elizabeth McKee Williams
Information Resources Specialist
University of Michigan


We carried out a survey of 230 science faculty and researchers at the University of Michigan in the departments of astronomy, biology, chemistry, geology, mathematics, natural resources, physics, and statistics, seeking to learn about their information resource use and preferences in diverse research contexts (e.g., current awareness, routine day-to-day information needs). The response rate was 46.5%. Results indicate that researchers rely largely on the same information-seeking tools in different research contexts. We also asked about the importance of six factors -- speed, convenience, familiarity, currency, authoritativeness, reliable availability -- on their information resource preferences; convenience and time were most frequently cited. Finally, open-ended questions invited respondents to comment on barriers and frustrations they encountered in information-seeking; themes of availability and access to needed material were mentioned, in addition to a wide variety of information retrieval difficulties.

Purpose of Study

This research project grew out of the University of Michigan Shapiro Science Library's planning session in the winter of 2000, with the recognition that we needed better information about the information services needs of the U-M's science faculty and research community. Although we believed that we had a good sense of students' needs, through daily interactions at the reference desk and through instruction sessions, we recognized that in recent years there have been important shifts in the ways in which science faculty and researchers at the university were seeking and using information. We became aware that the library needed to be more active in attempting to understand these changes.

Scientists' information use at the University (and, presumably, at other campuses) has been affected by several factors including:

The primary goal of this research project was to gather information about science faculty needs in conducting research -- what their priorities are, and in particular where they were encountering bottlenecks and frustrations. We formulated three desired outcomes from the project: first, to establish a baseline of information from which to develop reference services, as described above; second, to engage faculty and researchers in a dialog with the science library, and third, to develop an information-gathering tool that could be adapted and used again to gather information in future years.

In developing our set of research questions, we chose to look at the U-M science faculty as a community of individuals involved in producing and consuming information, in a variety of research-related contexts, with each person making choices about information-seeking strategies with their relative costs and benefits. We wished to establish an overall accounting of the types of information resource judged most useful in various contexts, and to discover what factors influenced scientists' choice of these resources. Finally, we wished broadly to understand what barriers or frustrations scientists encountered in conducting their research, and whether they felt their information seeking activities to be generally successful.


In selecting a survey pool, we started with a database of faculty in the departments served by the Shapiro Science Library. The database included all faculty, primary researchers, visiting faculty, fellows, and postdoctoral fellows in these departments. Excluded were emeritus professors, adjunct faculty and graduate students, although we may extend the survey to graduate students in selected departments at a later date.

The science departments, thus defined, ranged in size from 24 (Astronomy and Statistics) to over 100 (Mathematics and Chemistry). We decided to survey the two smallest departments, Astronomy and Statistics, in their entirety, in order to ensure that these small communities had a better chance of being represented. We then selected a stratified random sample, using SPSS software, of 42% from each of the other departments, for a total survey pool of 230 people. This percentage was chosen because our target response rate was 40%, which would give us nearly 100 survey responses.

Influenced by our research into the literature (Orr 1970, Brown 1999), we concluded that any inquiry into information use should take account of the context and nature of the need, and not evaluate information seeking in a vacuum. A person may make different choices, we felt, when doing routine, day-to-day information searches than when exploring an unfamiliar topic or area. Thus we identified four information-seeking contexts: 1) current awareness, 2) routine research needs, 3) a thorough search of the literature, and 4) a new or unfamiliar research area. These four contexts appear in each of the first two sections of the survey.

In the first section, we asked scientists to identify the resources that were most useful to them; they were permitted to select as many resources as they saw fit. This section was intended to elicit a straightforward accounting of the most useful types of resource, both formal and informal, in the different disciplines. We also wanted to see if scientists turned to different resources for different types of research need.

We were also interested in the factors that influenced a scientist in preferring one type of resource over another when faced with an information need. In his work analyzing the flow of information among scientists, Orr (1970) writes about information seeking as a rational decision-making process, with different costs and benefits associated with each information transaction. He suggests a model for examining information-seeking behavior using time as a measure of value -- how scientists choose to allocate their time in seeking information. This model formed the basis for this section of the survey. We presented six positive attributes [e.g., most convenient, most current] and, for each of the four information seeking contexts described above, we asked which three factors most influenced them in preferring one information resource over another.

The six factors were as follows:

The survey included several open-ended questions, a strategy that we believe helped provide balance to the conclusions we drew from the closed or multiple-choice responses. We asked the following four open-ended questions:

In analyzing the comments received in response to these questions, we first scanned them to identify broad thematic categories. Two of our group then worked independently to assign category codes to the comments; the results were compared and discrepancies reconciled. In most cases, agreement between the two coders was good; for one especially far-ranging question (the one about comparing other libraries to the U-M's library), the category definitions were revised twice until greater consistency between the two coders was reached. Each comment was given as many category codes as seemed appropriate, a necessary step since many comments addressed several issues at once.

The survey (see appendix) was pre-tested on three science faculty members during development. All questions were developed by our research group, with the exception of three questions about the impact on researchers of late discoveries of information, which were based on a 1987 British Library study (Martyn 1987).

Three rounds of responses were solicited by personally addressed e-mail messages; the first on July 14-18, the second on August 14-15, and the final round on September 8. Researchers were invited to complete the survey electronically by visiting a web site where the survey form had been posted. They were given assigned ID numbers so that they might have the option of anonymity in their responses. They were also given the option of requesting to be removed from the survey entirely; there were 24 such requests (10.4% of the survey pool).

The survey received 104 valid responses, for an overall response rate of 46.5%. There were three blank (invalid) responses, which were not included in the analysis but were counted towards the response rate. Of the 104 respondents, one submitted a second (slightly different) response; the two were merged so that all his/her choices and remarks would be included.

Results and Discussion

Most Useful Information Resources

This set of questions was intended to discover what types of resources were highly valued by researchers, and whether researchers' approaches to information seeking varied greatly depending on the context of their research need. Our hope was that library subject specialists would obtain useful information in these responses on which to base collection decisions and to aid in developing relevant and useful information services. The response rate for this section was 44.7%.

Responses indicate that science faculty and researchers at U-M rely on much the same types of information resource in different research contexts. Looking at the "top four" information-seeking approaches for each research context -- that is, the four types of resources most frequently cited as "most useful" -- two appear in all four contexts, and a third appears in three of the four contexts. The two that appear across all four contexts are "following leads" or a "citation trail," and reading review articles or journals. The one that appears across three out of the four research contexts (all of them with the exception of current awareness), is searching databases, indexes/abstracts, or library catalogs; this category of resource is the approach most frequently cited as useful in those three contexts. This result is noteworthy in light of indications in a later section of the survey that a significant number of respondents experience some difficulty and/or some degree of frustration with the use of formal information retrieval/discovery tools (see Section C, Responses to Open-Ended Questions).

In each information-seeking context, one or more types of resource were preferred that did not appear in the top four for any other context. For example, in managing routine information needs, consulting a personal collection of books, journals, or photocopied articles, was the third most frequently selected resource; and in seeking information in a new or less familiar area, consulting a colleague was the fourth most frequently chosen resource. These approaches were not among the top four in any other research context.

Though we found consistency in the types of resources most valued by researchers across different research contexts, there was considerable variability in the types of resources considered useful across different disciplines. In some cases this may simply reflect availability. An example is preprint literature, which is a mainstay of research in astronomy and mathematics. Preprints were considered very useful in staying abreast of current research developments by 70% of astronomers, 78.9% of mathematicians, but only 9.1% of chemists and 13.3% of geologists. In the latter disciplines, preprint literature is not only less available, but in some cases is actively discouraged (e.g., the American Chemical Society journal editors' recently announced policy that preprints constitute prior publication and thus render an article ineligible for publication in ACS journals, "Preprint Policy" 2001).

Formal current awareness tools and alerting services, such as Current Contents, were deemed useful by only 30.8% of the respondents overall, ranging from lows of 0% (Statistics) and 10% (Astronomy -- which, as mentioned above, relies heavily on the LANL preprint server) to highs of 63.6% (Chemistry) and 66.7% (Natural Resources & Environment). It is unclear whether the relatively low use of these services among some disciplines stems from the emergence and popularity of alternatives (preprints, previews, and the like) or from a judgment that these tools are unsatisfactory in some way, or perhaps a low level of awareness of the existence and utility of this service. Evidence for the relatively low level of awareness among some faculty of library-provided tools and services is further discussed in Section C, Responses to Open-Ended Questions, and supports findings in other studies, notably the 1997 study at UC-Berkeley (Maughan 1999).

Table 1. Top Four Information Resources Judged Most Useful by U-M Science Faculty and Researchers... Current Research Developments

Four Most Frequently Selected Resources:
  • Browse recent issues of print journals (67.3% of respondents)
  • Attend conferences, symposia, or other professional meetings (66.3% of respondents)
  • Follow references or leads form an article or item of interest: "citation trail" (54.0% of respondents)
  • Read review journal(s) or review article(s) (49.0% of respondents)

...for managing Routine Research Information Needs

Four Most Frequently Selected Resources:

  • Search databases, indexes/abstracts (e.g., PubMed, Biological Abstracts) or library catalogs (e.g., MCAT) (75.7% of respondents)
  • Follow references or leads from an article or item of interest; "citation trail" (71.8% of respondents)
  • Consult personal collection of books, journals, photocopied articles, reprints, etc. (54.4% of respondents)
  • Read review journal(s) or review article(s) (53.4% of respondents)

...for conducting a Thorough Search of the Literature

Four Most Frequently Selected Resources:
  • Search databases, indexes/abstracts (e.g., PubMed, Biological Abstracts..) or library catalogs (e.g., MCAT) (80.4% of respondents)
  • Follow references or leads from an article or item of interest; "citation trail" (70.6% of respondents)
  • Read review journal(s) or review article(s) (63.7% of respondents)
  • Browse/scan publications, including books, journals, indexes/abstracts (43.1% of respondents)

for finding information in a New, or Less Familiar Area

Four Most Frequently Selected Resources:
  • Search databases, indexes/abstracts (e.g., PubMed, Biological Abstracts..) or library catalogs (e.g., MCAT) (68.6% of respondents)
  • Follow references or leads from an article or item of interest; "citation trail" (58.8% of respondents)
  • Read review journal(s) or review article(s) (55.9% of respondents)
  • Consult a colleague (e-mail, telephone, mail, or in person) (48.0% of respondents)

Preference Factors in Choosing a Research Resource

This section of the survey was intended to gain insight into the factors that influence science researchers in choosing one research resource over another. The response rate for this section was 44.0%.

Of the six positive attributes that were presented, "Most Convenient" was the most frequently cited factor in choosing a research resource, in three out of four contexts (for current information seeking (76.5% of respondents), for routine research information needs (77.5%), and when seeking information in a new or unfamiliar area (59.4%)). In each of these three contexts, "Least time..." was the second most cited factor influencing the choice of a particular information resource (for current information seeking (55.9% of respondents), for routine research information needs (61.8%), and when seeking information in a new or unfamiliar area (50.5%)). In the context of conducting a thorough search of the literature on a particular research topic, by contrast, "Most familiar" was the most frequently cited preference factor (68.0% of respondents).

Authoritativeness did not appear to be a factor of particular importance to scientists in selecting an information resource in most research contexts, perhaps because they rely instead on their expertise in assessing the value of the information they find. Authoritativeness may be a more relevant and familiar consideration for librarians than for scientists. In the context of conducting a thorough search of the literature, "Most authoritative" was not selected by any respondents as a factor most influencing their choice of an information resource. In seeking information in new or unfamiliar areas, authoritativeness was cited as an important factor by 46.5% of respondents.

Responses to Open-Ended Questions About Information Use

The open-ended questions were intended to elicit information about barriers and frustrations in information-seeking that we might otherwise be unaware of. Because of the broad nature of many of the thematic categories that we developed in coding the responses (e.g., "Unavailability: material not on shelf, in the remote storage facility, in disorder, missing, not owned/licensed by U-M, access restrictions"), and because of the inevitable overlap of some of the categories, and the differences in degree of specificity among them, we tended to be conservative in interpreting the results. We made the following general observations.

Science researchers at U-M perceive information retrieval difficulties (such as inadequate discovery/retrieval mechanisms) to be a barrier in seeking research-related information, along with lack of time, and the unavailability of needed material. This thematic category, information retrieval difficulties, grouped together a broad range of issues, including confusing search systems, or confusing or inadequate indexing of material, as well as the user's own lack of skill or comfort with search systems or with technology in general. Comments here included remarks about the lack of electronic access to older literature, "grey" literature, and foreign literature; poor control of government and state publications; and inadequate coverage by indexes of journals in certain areas. Also noted were frustrations with the vagaries of online searching: "all too haphazard for my tastes," "poor indexing of topics," and "choosing the correct keywords" were mentioned. "Improvements in information retrieval/discovery mechanisms", was a desired change mentioned by relatively few (12.5%) of the respondents.

The top three changes wished for by the respondents (with equal frequency) were improvements in the availability or accessibility of material; more journals available online; and the removal of various procedural and technical barriers, such as those that occasionally impede photocopy, circulation, and ILL transactions. A number of these suggestions related to services that are in fact already available to U-M faculty and researchers. Examples include a service to deliver books and photocopied articles to faculty offices (a service called 7-FAST); renewing books online; e-mailing circulation notices rather than hard-copy mailing; and a service to get books delivered from one library unit to another. Faculty's relative lack of awareness of available library services is a theme also noted by Maughan (1999).

One area of frustration surfaced in each of the three questions that asked explicitly about perceived barriers and wished-for changes. This broad category embraced issues having to do with the availability or accessibility of material. Several respondents specifically noted frustrations associated with materials that are missing, misshelved, checked out, or in remote storage, adding that it was important for material to be available in practice and not just in theory. Resources not available online, or with restricted access (limited number of seats) was a specific instance of unavailability mentioned by a few. In thinking of other libraries and information providers that they had had experience with, several respondents commented favorably on libraries where collections were physically close at hand, both departmental libraries and those in which older materials were immediately accessible and not in remote storage.

Another broad theme that emerged across all four questions dealt with frustrations experienced by science faculty and researchers as a result of the consolidation of the Science Library from several departmental libraries in 1995. These concerns were typically expressed in terms of time and (in)convenience -- the two preference factors that emerged as top priorities for researchers in section two of the survey. The time pressures that commonly afflict academic researchers are especially acute in many science disciplines due to their rapidly advancing research fronts, and the lack of time to acquire, filter, and digest information was a concern explicitly and eloquently noted by several respondents.

Some respondents specifically commented on the time it takes to visit the library because it is physically distant from their departments; others noted that their work requires that they visit several libraries on campus, hence a lot of "running around." Some scientists whose research areas are interdisciplinary, on the other hand, expressed frustration with the distribution of materials across various library units on campus (e.g., the engineering, medical, and museums libraries).

The most frequently mentioned themes emerging from these four questions, arranged by category, appear in Table 2.

Table 2: Answers to Open-Ended Questions

What is the greatest barrier you typically encounter in seeking information for your research?

85 Respondents answered this question. Responses were grouped into 9 categories; each response was assigned to as many categories as was appropriate.

Most Frequently Mentioned Themes
  • Information retrieval difficulties 29.4% of respondents (n=25).
  • Time 23.5% of respondents (n=20).
  • Unavailability 20.0% of respondents (n=17).
  • Inconvenience (physical location of library, multiple library locations, effort) 15.3% of respondents (n=13).
No Response: n=19

What do you enjoy most about seeking information?

77 respondents answered this question. Responses were grouped into 4 categories; each response was assigned to as many categories as was appropriate.

Most Frequently Mentioned Themes
  • The thrill of the hunt. 51.9% of respondents (n=40).
  • Advanced capabilities of information technologies/convenience. 31.2% of respondents (n=24).
  • Finding unexpected information; serendipity 16.9% of respondents (n=13).
  • Several people remarked on the experience of finding information that was not originally sought; little surprises and unexpected connections.
No response: n=27

Which libraries or information services have you used (either physically or remotely) that you might compare to the University of Michigan's library? What features or services did you particularly like, or dislike?

53 respondents answered this question. Responses were grouped into 8 categories, representing the type of feature or service that was discussed; each response was assigned to as many categories as was appropriate.

Most Frequently Mentioned Themes
  • Availability of materials (collections, access, organization) 37.7% of respondents (n=20)
  • Specific electronic resources (whether commercial or freely available) 26.4% of respondents (n=14)
  • Physical location/convenience of library 24.5% of respondents (n=13).
  • Procedures, services, facilities (e.g., circulation, photocopying) 20.8% of respondents (n=11)
No response: n=51

If you could change one thing that affects your interaction with the library or other information providers, what would it be?

64 Respondents answered this question. Responses were grouped into 8 categories; each response was assigned to as many categories as was appropriate.

Most Frequently Mentioned Themes
  • Availability of Material. More accessible, owned/licensed, available, browseable, orderly. 21.9% of respondents (n=14).
  • More Material Online. 21.9% of respondents (n=14)
  • Procedural/Technological Barriers Removed. Improved services. 21.9% of respondents (n=14)
  • Saving Effort. Convenience, in particular the physical location of the library. 18.8% of respondents (n=12)
No Response: n=40

Conclusion and Next Steps

We felt that the better-than-expected response rate, and the thoughtful nature of many of the comments offered evidence that two of our secondary desired outcomes have been accomplished-the furthering of a dialog with science faculty, and the development of a survey instrument that can be adapted for future use by the library. While it is not our claim that these findings and conclusions can be extrapolated to the entire scientific community, we believe that they will be helpful to us in understanding the concerns of our local community of researchers, and may be of interest to other academic science librarians.

We felt that the hybrid nature of the study worked well. Closed, check-box type questions yielded quantitative data that was often reinforced by the conclusions drawn from analysis of the comments. The open-ended questions protected us from blind spots that we might otherwise have missed, like the concern for availability and access to physical materials that was voiced by many in the comments section, but which we hadn't asked about in any other way.

The study has suggested several possible avenues for future action, ranging from simple and immediate to more complex:

A number of the services that scientists identified as desired changes are, in fact, already available to them. Examples are the delivery of materials from one library unit to another, or the ability to issue a photocopy card tied to a research fund. It should be a goal to explore new and more effective methods of promoting existing services, and facilitating their adoption.

Specific comments about issues of access and availability of materials (e.g., circulation of bound journals in particular disciplines) can be addressed. As noted in the 1997 UC-Berkeley study, any steps towards addressing these issues cannot help but improve end-user satisfaction (Maughan 1999). Solutions such as a special project to catch up with a binding backlog in the current journals room will help in keeping current journals orderly and accessible. The Science Library has also recently adopted a policy allowing for the occasional circulation of bound journal volumes, under certain limited circumstances, for faculty and graduate students in order to accommodate particular needs.

As a result of the study, library subject specialists have gained additional insight into the way research is conducted in particular disciplines -- in particular, which resources and information-gathering strategies are effective and highly valued, and which are not.

The recognition that there is a certain level of frustration with information retrieval mechanisms is a problem that is less amenable to an easy solution. This may be an area where science librarians can work to get more detailed feedback on specific information systems, and to provide more relevant support for searching and information retrieval.

One theme, which emerged across all four questions, dealt with frustrations experienced by science faculty and researchers as a result of the centralized Science Library. Some also expressed frustration with the distribution of materials across various library units (e.g., the science, engineering, medical and museums libraries). While centralization does allow the library to provide improved access to a wide array of resources and services, we recognize the barriers that arise for many as a result of the centralized library system, and we will continue to explore best ways to bring library resources to researchers when and where they are needed.

Appendix: Science Faculty Information Needs Survey


Brown, C. M. 1999. Information Seeking Behavior of Scientists in the Electronic Information Age: Astronomers, Chemists, Mathematicians, and Physicists. Journal of the American Society for Information Science 50(10): 929-43.

Crawford, S. Y., Hurd, J. M., & Weller, A. C. 1996. From Print to Electronic: The Transformation of Scientific Communication. Medford, NJ: American Society for Information Science.

Martyn, J. 1987. Literature Searching Habits and Attitudes of Research Scientists. British Library Research Paper no. 14. London: British Library Research and Development Department.

Maughan, P. D. 1999. Library Resources and Services: A Cross-Disciplinary Survey of Faculty and Graduate Student Use and Satisfaction. Journal of Academic Librarianship 25(5): 354-366.

Orr, R. H. 1970. The Scientist as an Information Processor: A Conceptual Model Illustrated with Data on Variables Related to Library Utilization. In C. E. Nelson & D. K. Pollock (Eds.), Communication Among Scientists and Engineers. Lexington, Mass.: Heath Lexington.

Preprint Policy. 2001. Chemical & Engineering News 79(3).

Smith, E. 1990. The Librarian, the Scholar, and the Future of the Research Library. New York, NY: Greenwood.

Soper, M. E. 1976. Characteristics and Use of Personal Collections. Library Quarterly 46(4): 397-415.

Von Seggern, M. 1995. Scientists, Information Seeking, and Reference Services. The Reference Librarian. 49-50:95-104.

Previous   Contents   Next

W3C 4.0