Carl Westin

50 First empirical insights 3-3-7 Dependent measures Dependent measures included: • Acceptance of an advisory (binary, accept or reject). • Agreement with an advisory (on a 1-100 scale). • Response time (from advisory onset to accept or reject button press). • Subjective scenario difficulty (on a similar 1-100 scale). 3-3-8 Procedures The entire simulation lasted four weeks. In the first week, the prequel simulation was conducted to capture participants’ solutions to the designed conflicts. Follow- ing briefing and consent procedures, we conducted sixteen training runs and sixteen measurement runs. The conformance analysis and programming was carried out during the middle two weeks. The dataset with manual conflict solutions was pro- cessed to create conformal and nonconformal resolution advisories for the experi- ment simulation. Eight conformal advisories (not necessarily eight different), and eight nonconformal advisories, were created for each participant. In the final week, the same participants again took part, this time by interacting with sixteen automated aided scenarios. Participants performed the same task as in the earlier prequel simulation, but now assisted by an automated aid (presented as a midterm strategic aid called the “automation advisory”) that would provide resolution advisories by proactively auto-selecting a conflict aircraft. These indi- vidually tailored advisories consisted of the eight conformal (again, unrecognizable replay) and eight nonconformal (again, a colleague’s different solution to the same scenario) resolutions. The essential subterfuge in our study (i.e., that “automation” was in fact only a replay) required participants to enter the experiment simulation entirely naive. Each participant was, therefore, carefully instructed during the de- brief to avoid discussing this topic with colleagues. Figure 3-5 shows the simulation interface, as it appeared in the experiment sim- ulation. The resolution advisory itself consisted of a heading vector, a speed vector, or a combination thereof. A resolution advisory was automatically presented in the SSD display of the aircraft. The suggested solution was depicted in one out of three ways: an orange vector line (heading advise), an orange ring with a di- ameter greater or less than the current speed as indicated by the green ring (speed advise), or a combination of both (heading and speed advise). The resolution ad- visory was accompanied by a beeping sound and a conflict dialog window (upper right corner in Figure 3-5) that the participant used to either ‘accept’ or ‘reject’ the