Date: August 9, 2012
Creator: Harker, Karen
Description: This presentation is for a workshop and discusses full-text linking problems. Full-text linking could be considered in much the same way as Mark Twain considered the weather - everybody talks about it, but nobody does anything about it. Many administrators of full-text linking services hear the general complaints that the service doesn't work, but few actually report specific problems that could be tracked and resolved. How can you, as administrator, validate these vague and general concerns? The first step could be simply to determine the extent of the problem. How bad is it? This seems simple enough - pick a few links, count how many get to the full-text and report this result. But how many links? From which sources? To which targets? Using a scientifically-sound methodology that randomly selects links, takes such factors into consideration, and uses statistical techniques to analyze the results, you can feel confident that your tests of the system are valid and reliable. This hands-on workshop will present this method, providing background to basic scientific and statistical methods. Participants will be able to try-out different methods of selecting a sample, determining the sample size, defining the possible outcomes, testing the links, and presenting the outcomes, ...
Contributing Partner: UNT Libraries