Abstract
Objective – To investigate and compare the nature of e-mail reference services in academic and public libraries outside the United States.
Design – Longitudinal comparative study.
Setting – A total of 23 academic and public libraries in ten countries: Australia, Belgium, France, Germany, Mexico, the Netherlands, Slovenia, South Africa, Sweden, and the United Kingdom.
Subjects – The authors collected reference questions that were e-mailed to the 23 libraries for the weeks of April 3, 2006 and April 7, 2008. Questions were sent from the libraries’ websites to QuestionPoint, a collaborative, online reference service that was used to answer the questions received.
Methods – The authors randomly selected 25 questions for each library for the weeks under investigation. If a library did not receive 25 email reference questions that week, then they collected transactions from subsequent weeks until the quota was met or until the end of the month. The authors examined transactions from a total of 919 questions – 515 questions in 2006 and 476 in 2008. All identifying information about the user was stripped from each transaction collected. Each transaction was labeled according to the following categories:
• Type of institution, i.e., whether the question was sent to an academic or public library
• Language of the question
• Question type, i.e., whether the question was about library policy or access to electronic resources (labeled “access” questions), about library holdings (labeled “bibliographic” questions), or about finding specific information on a topic (labeled “subject” questions)
• Answer type, i.e., whether the response consisted of: a confirmation, clarification, fact, instructions, referral to a pathfinder/bibliography, referral to another library/person/place, or no answer.
• User status, i.e., whether the person asking the question was an undergrad, a graduate student, or a staff/faculty member
• Subject classification of the questions using the Dewey Decimal Classification system
• Response time
Main Results – The e-mail transactions that were examined revealed a wide range of end-user and librarian behaviors. English, followed by Dutch, German, and French, were the languages most frequently used by library users. Countries also varied in terms of the types of questions received. For example, more than 75% of the email queries in Belgium (which only had academic libraries participate in this study) were “access” questions, while Mexico (which also consisted of all academic libraries) only received 6% “access” questions, France (all public libraries) had relatively few access questions, and Sweden (also all public libraries) had none. Public libraries received the most “subject” questions (75%) compared to academic libraries (28%). Public libraries answered “subject” questions with facts over a third of the time, while academic libraries responded with instructions close to half of the time.
Among the academic libraries, graduate students asked slightly more “access” questions than undergraduates (62% versus 56%), and undergraduates asked more “subject” questions than graduate students (26% versus 13%). The “subject” questions submitted to academic libraries were divided almost equally among topics in the humanities (36%), the sciences (32%), and the social sciences (32%). This differed from public libraries; the latter received mostly questions about humanities topics (65%).
The time taken to respond to users’ reference questions ranged from a few minutes to a few weeks between libraries. Some libraries set the response times on their websites. Those libraries that indicated longer response times on their sites met the users’ expectations more often, up to a maximum of 100 percent of the time.
Most of the characteristics of email reference services that are listed above remained consistent from 2006 to 2008. The two areas that changed over two years were the libraries’ response time and the types of questions asked by university students. “Access questions increased (by 14 percent among graduates and by 4 percent among undergraduates), and bibliographic and subject questions decreased in both groups” (p. 364). Response time improved overall from 2006 to 2008.
Conclusion – The authors’ analysis of the 919 transactions of e-mail reference questions revealed unique patterns about end-user information seeking behavior and librarians’ responses in academic and public libraries outside the United States and Canada. One of these patterns is that the public libraries participating in the study received the highest percentage of “subject” questions. The authors state that “the pattern of a much higher percentage of subject-related questions in public libraries contrasts with the general virtual reference trend in academic libraries, which shows a much higher percentage of access questions. Since many of the access questions concerned connection problems or logging on to databases, the relatively fewer number may indicate that the arts and humanities disciplines require less database searching and that the users need specific answers instead” (p. 367).
The data also revealed significant differences between the types of questions asked by undergraduates versus graduate students. Undergraduates asked two thirds of the subject questions submitted to academic libraries and graduate students asked just over a fourth. The authors assume that this finding indicates that graduate students do more of their own research than undergraduates.
The authors were concerned by the increase in the number of access questions posed by undergrads and graduate students from 2006 to 2008. They suggested that websites, databases, and other resources might have become more difficult to use over the years. They also noted that questions in technology almost doubled from 2006 to 2008.
One of the patterns that were revealed contradicted the authors’ assumption that libraries with slow response times in 2006 would improve in 2008 as they became more proficient in providing virtual reference services. The majority of libraries in the study improved their turnaround time from 2006 to 2008, but the two slowest libraries took even longer to respond to their users.
Comparing different response time threshold setting methods to detect low effort on a large-scale assessment