Conclusion

5.1  Impact for practitioners

Our study, by its high-specificity nature, holds little real-world value. However, useful tangents and conjectures may be gathered from aspects of the experiment. Of interest is subjective satisfaction and user remarks. Some subjects mentioned that they looked only at the query devices while ignoring the starfield display. What good is a graphical query interface such as Spotfire if users don't look at the visual element? Certainly, users with real motivation would use the software as intended. So a question for both researchers and practitioners arises; how do you make certain that users will correctly use your DQI correctly?

In our observation of subjects while they worked through the test trials we saw that users, given a fairly comprehensive tutorial on most of Spotfire's features, only used a few of these features to perform their tasks. As noted before, some subjects didn't bother to look at what the starfield display showed them, thus they proceeded to further filter out data when it was not necessary. This suggests that while dynamic query interfaces are a powerful tool, it is at first only a potentially powerful tool. DQI users must be trained on how to realize the benefits of direct manipulation and DQI.

5.2  Suggestions for future researchers

  • Perform multiple trial runs of subjects with different orderings of test cases
  • Use different query requirements for each task, instead of re-using the same task(which unfortunately allowed our subjects to memorize or familiarize themselves with the tasks, creating a possible bias).
  • Attempt to apply useful real-world findings based on experiment results.
  • Repeat experiment with expert users of Spotfire or other DQI, effectively removing "learning" bias.
  • Discover how to get users to utilize all functions of the DQI, including query devices and graphical displays(we found that many users did not look at what the starfield display showed them).

    Return to Main
    Return to SHORE '98