The idea about the eye tracking study was born in July 2010 when Natalia Konevtsova started her diploma thesis in our VISUS institute. She was very interested in designing controlled user experiments and liked working with participants. Her diploma thesis was more than just about this single eye tracking experiment. She wrote about the different phases that have to be taken into account when conducting such time-consuming experiments. In her thesis she explored what is important before, during, and after a controlled user study.
The final result of this eye tracking study was a very interesting insight: When asking participants about finding out the least common ancestor of a number of red colored and marked/highlighted leaf nodes in traditional, orthogonal, and radial node-link tree diagrams, participants are much faster in the non-radial (traditional and orthogonal) diagrams. Furthermore, they used some kind of cross-checking behavior to explore the radial tree diagrams, a phenomenon that did not occur that often in the non-radial diagrams. In our study, we also switched the tree orientations for the non-radial diagrams. Here, we uncovered that the diagrams with the root on top performed best and are liked the most by the participants. There are very many insights, some of them are presented in our corresponding VisWeek 2011 Paper. A more sophisticated statistical analysis of the gaze trajectory data will follow soon.
The final message of the experiment is: "Use traditional node-link tree diagrams with the root oriented to the top (top-down diagram)!"
If you are interested in our study, read the corresponding VisWeek 2011 paper, or write an email. I will answer your questions as soon as possible.