Carl Westin

5-5 Method 101 Two questionnaires were used to assess differences in participants’ perceptions of automation transparency, and in particular their understanding of the two inter- face representations. The first questionnaire consisted of two identical sets of eleven Likert statements, one for each SSD representation (HB and TRI), with a seven- point scale ranging from strongly disagree to strongly agree. Most of the statements were adapted from transparency questionnaires in Cramer et al. 111 and previous studies of our own. 101, 217 The second questionnaire consisted of a seven questions on a Visual Analogous Scale (VAS). The VAS is a subjective questionnaire instrument on which responses to statements or questions are made by making a mark somewhere on a 100 mm continuous line that connects two contrasting endpoints, with a marking in the mid- dle indicating a neutral response. Here, endpoints consisted of the two representa- tions encountered in the simulator. Questions addressed participants’ representation preference in relation to how well each representation facilitated understanding of why a certain advisory was suggested (i.e., why transparency), and how well the constraints of the conflict and its alternative solutions were visualized (i.e., domain and tradeoff transparency). Additional questions addressed the representations use- fulness for CD&R, the difficulty experienced in working with them, the perceived clutter, and the perceived workload. 5-5-7 Procedures The three-phased study ran over a three-week period and encompassed two simula- tions. Both simulations were preceded by consent procedure, briefings, and training runs. The initial prequel simulation, conducted in the first week, captured partic- ipants’ conflict solutions of the designed conflict. The simulation consisted of ten two-minute en-route scenarios. Unknowingly to participants, the measurement sce- nario containing the designed conflict was repeated four times. In this phase partic- ipants solved conflicts using only the HB SSD. In the second week ( conformance design phase), participants’ four recorded solutions were analyzed to define their individual solution style, consisting of the most persistent pattern found. The individual conflict solutions style determined each participant’s conformal and nonconformal resolution advisory to be used in the subsequent main simulation. For example, consider the conflict between aircraft A and B in Figure 5-4. If Participant 1 repeatedly solved the conflict by vectoring aircraft A to the right behind aircraft B , the conformal solution style would be to vector aircraft A to the right behind aircraft B . A nonconformal, opposite solution style, could then be that of Participant 2, who consistently vectored aircraft B to the left behind aircraft A .