Abstract: Crew Resource Management (CRM) and Threat and Error Management (TEM) are known to be reliable techniques to improve human performance by preventing and avoiding flight errors in cockpit. Neither CRM nor TEM, however, has carried out an in-depth systematic analysis on the error-inducing navigation behavior and, as a result, there is a lack of reliable behavior objectives or effective evaluation measures for implementing trainings on error preventing skills, problem diagnosis on navigation behaviors, and effective evaluation on safety performance. Although a relatively comprehensive evaluation system for technical-skill assessment on airline pilots has already been adopted in some countries, the system for evaluating non-technical skills remains weak. The current situation characterized by more emphasis on examining technical skills while little on evaluating and training non-technical skills/human factor skills leads to one of the major reasons for the lasting high accident rates induced by human factors in current aviation safety field. The fundamental key to this issue, necessarily, is to start with evaluating and diagnosing the standardization of navigation safety behaviors of airline pilots. Unlike other industries adopting the rate of accidents and incidents or injury rate as safety index, the airline industry has adopted “the standardization of navigation behaviors of airline pilots” as a fairly effective index to evaluate safety operation performance, and diagnose operating errors, due to its character of being able to objectively and veritably capture various potential risk behaviors leading to flight accidents or incidents. On the basis of combining CRM and TEM models with the characteristics of Chinese aviation safety culture, the present study established a multidimensional evaluation method to assess navigation safety behaviors of Chinese airline pilots, aiming at providing a set of objective and effective assessing instruments for selecting and training Chinese airline pilots, and evaluating and diagnosing the standardization of navigation behaviors in airline flight.
All the 118 subjects came from China Southern Airlines Ltd. as college graduates with good physical and mental health. They met physical examination standards formulated by Civil Aviation Administration of China. Through literature review, expert interview, and open-ended questionnaire survey, twenty-seven representative aircrew behavior patterns which could influence modern airline flight safety were selected, which, in turn, composed the test items for assessing navigation safety behaviors under Chinese culture background. Four flight experts were invited to conduct Level Evaluation on the actual flight performance of the 118 airline pilots, according to the test content throughout the following four flight phrases: preparation/skidding, taking off/climbing, cruise, and descending/approach/landing — from "Lower than the Anticipated level (1)" to "Crew Paragon (4)".
The results of Confirmatory Factor Analysis (CFA) conducted on the evaluation data concerning the navigation safety behaviors of 118 Chinese airline pilots indicated that the four-dimension model composed of Automation System Understanding, Leadership and Management, Situation Awareness and Decision-making, and Interpersonal Communication and Cooperation was the best fit of data among the 5 hypothesized models. Moreover, it is of satisfying reliabilities and construct validities throughout the four flight phases. With safety performance accessing data of 86 pilots (in terms of Safety Regulations, Flight Style, Flight Skills and Aircrew Management) as criterion variables, and the pilots’ flight performance data throughout the different phases of airline flight as predictive variables, a series of hierarchical regression analyses were implemented, respectively. The results showed that the present multidimensional scale had high differential validity and predictive validity; the four dimension variables contributed significantly to the evaluation of navigation safety behaviors in different flight phases, with differentiated contribution levels according to different mission requirements in varied flight phases.
In sum, the multidimensional scale developed in this study not only provides an effective tool to standardize the navigation safety behaviors of crewmembers, to identify the errors in aircrew navigation behaviors, and to examine safety performance, but also lays a solid theoretical foundation for developing programs specifically for pilot selection and aircrew CRM trainings.

Key words: airline flight, non-technical skills, safety behaviors, assessment, multidimensional scale