Document Type


Publication Date





When independent variables are inter-correlated with each other, ANOVAs or traditional multiple regression methods do a poor job for analyzing their relative importance in accounting for the variance in a dependent variable. This paper describes a method called Dominance Analysis (Budescu, 1993; Azen & Budescu, 2003) as a better approach than the traditional methods in determining the relative importance of several inter-correlated independent variables in accounting for the variances in pilots’ performance measures in a conflict detection task with a cockpit display of traffic information (Xu, Rantanen, & Wickens, 2004). The three variables in question were an intruder aircraft’s distance to closest point of approach (CPA) between the pilot’s ownship and the intruder aircraft, the intruder aircraft’s time to CPA, and relative speed between the two aircraft. Results indicate (1) for absolute miss distance estimate error, distance to CPA was the most important variable than the other two variables; (2) for signed miss distance estimate error, time to CPA and distance to CPA were more important than relative speed; (3) for both absolute and signed time to CPA estimate errors, time to CPA was the most important compared to the other two variables; and (4) for absolute orientation at CPA estimate error, relative speed was the least important variable compared to distance to CPA and time to CPA. Interpretations of the dominance analysis results are offered.