Barbara M. Wildemuth, profesorka z University of North Carolina at Chapel Hill ze Spojených států amerických, vystoupila na konferenci Inforum 2003 a příspěvkem nazvaným "Why conduct user studies? The role of empirical evidence in improving the practice of librarianship". Během konference jsme s ní připravili rozhovor, který autorizovaný uvádíme v původním znění bez titulků.

Could you briefly introduce your institute, the School of Information and Library Science at the University of North Carolina?

The School was founded in 1931 and has provided library science education for its entire history. We started a PhD program in 70s. In 1987 the School´s name was changed to the School of Information and Library Science and we began to offer a master´s degree in information science. Most recently we began an undergraduate program in information science. We started with a minor, which means that students would have a major in one area and information science as their second area. This year, we launched an undergraduate major in information science. We have about 350 students, over 200 students in master´s programs, about 35 undergraduate students and about 50 doctoral students. The School has 20 faculty members; our dean is Joanne Marshall and the school has recently sponsored a trip to the Czech Republic (viz článek v tomto čísle "Sláva moravským knihovnám", pozn. PJ).

What is your position at the school? What are your areas of research and interests?

I am a regular faculty member, which means I have both teaching and research responsibilities. I teach a variety of courses. At the master´s level I teach a systems analysis course, which is required for all information science students and is also available to library science students. This fall, I also will be teaching a course on information ethics. For undergraduates, I have taught a course on retrieving and analyzing information. In my research, I mostly focus on how people interact with information, particularly through electronic media. I have a couple of different projects going on right now. One of them is the Open Video Project, a library of digital videos. It is "open" in the sense that the videos are available for free for the research and educational communities for video production. Our research with that is oriented towards development of a search engine which works with video material and an equivalent to an "abstract" in text materials, surrogates that describe video content. Just as you have abstracts to represent the document, and you read the abstract before you decide whether you want to read the whole document, we are developing such surrogates for videos, which will show you what the film is about, not just describe its content. The project leader, Gary Marchionini, is presenting a paper this week at the ACM-IEEE Joint Conference on Digital Libraries, which shows how fast we can make fast-forward surrogates since we investigated a variety of speeds.

How do you develop the content of the Open Video collection? Where do you find video materials?

The Open Video collection is made of contributions, mostly from organizations which have existing collections of videos that they would like to have digitized and made available. We have some videos from the Library of Congress; we have a number of videos from the Internet Archive, which is a large internet collection; we have some from the Human-Computer Interaction Lab at the University of Maryland and some computer-human interaction demos from conferences; we have some from the Informedia Project at Carnegie Mellon University; and we just got a series of films from the Digital Himalaya Project.

You presented paper on user studies at the conference INFORUM 2003. How is this subject connected with your teaching and research activities?

It is the main purpose of my research since I am focused on how people interact with information. I mainly study people while they are interacting with information. It is really easy for us as librarians to think that people interact with information in a certain way because it is similar to the way that we would interact with that information. In general, people who choose to be librarians have a certain relationship with information that is different than most other people. That is why we really need to study particular target audiences and how they interact with information, and then develop tools that will help them. Often, the outcome of my research is the design of such an interface. In my paper, I tried to encourage people here to conduct user studies in their own libraries, because I believe that even if they are not designing systems, they are in fact designing services for their users. Therefore they need to study the information needs and information behaviors of their target groups.

I am personally very glad that you introduced this issue to the Czech audience here at the conference. I think there is always this risk that even nicely designed library building with a rich collection of valuable materials and modern on-line catalog could be hardly understandable and useful for non-librarians. You called your research topic "evidence-based librarianship". Is this term your own "invention"?

No, in my paper I note that this idea has been discussed for several years by medical librarians, and it has been used earlier by Joanne Marshall. We both agree that this is a useful kind of a model. The idea of evidence-based medicine has been around for over 30 years and I believe it is a very good model, that over a period of time you can accumulate a number of studies looking at the same issue - not just one study- to start to understand that issue. Librarians can and should adopt this model.

Could you then recommend to Czech libraries where to start adopting this model? Do you have any "best practices" to share?

One example is a study that I am doing right now. It focuses on how people are conducting known-item searches in on-line catalogues. When you go to the library to do a known-item search, you already have some information about the item, either from somebody else, or from some citations, a reference list or something like that. We are capturing what the person has with them when they come, literally photocopying it, so we know exactly what they have. We are capturing the transaction logs so we know which pieces of that they actually used in their search. Then we ask them to show us the item they finally found. (We also replicate their searches, since they might not have found the item if they made search errors or if the library doesn´t hold the item. This way, we can distinguish collection failures from search errors.) From this data, we can make some suggestions about what should be put into on-line catalogues. On-line catalogues are a really sensitive issue, since librarians believe they have to describe everything about the book. But I am not really sure we do...

What do people actually have in their hands when they come to the library to search the on-line catalogue?

We did not analyze all our data yet, but I think it is title words that people use most often. What was really surprising to me was that what people brought in their hands was almost always accurate!

Human-computer interaction studies usually focuses on how a digital interface should be designed to better serve users, but do you also study how these systems actually modify human beings in a way that people are limited by using only those materials which are accessible?

For such a study we would probably need more than a decade of research since human behavior is changing very slowly. I have not seen any systematic study of this question. However, we did some studies with medical students, where we compared one class of the students in their first year of studies with the same group when they were in their third year. In both cases we were looking at search strategies. In the first year they were using a very small factual database about microbiology and in their third year they used MEDLINE. The outcomes from this research make me believe that it would be useful to look at people´s search behaviors and how they evolve. There are some experimental studies concerning the impact of aspects of Web search engine design. For instance, when we give people a small search box, they enter only few search terms, and they enter more terms when we give them a larger search box.

Producers and vendors are asking for feedback, but what they are getting at this point is feedback from librarians about what the librarians prefer. I think that librarians would be in a very powerful position if they did some user studies saying, for instance, "My users have problems with this feature, they do not understand how to use it, make it better!"

Could you tell us how the community around the conference here appears to you?

I found it very exciting to see how many people are interested in improving library services. It was also very nice to see how international the conference is; I met people from all the neighboring countries here. The conference is a very good mix of introductions to new products and services, presentation of research outcomes, and more analytical papers which provide a wider view of information issues.

Thank you for the interview.