航线驾驶安全行为多维评价量表的构建

结合现代航线驾驶工作特性及CRM和TEM模型基础上, 通过文献分析、专家访谈和开放式问卷调查, 筛选出现代航线飞行员驾驶安全所必备的27种关键行为特征, 构成中国文化背景下航线驾驶安全行为评价量表的检测项目。对118名中国航线飞行员测评数据的验证性因素分析表明, 由自动化系统认识、领导和管理、情境意识与决策、人际交流与合作技能组成的四维模型优于其他假设模型, 且在四个飞行阶段中均具有良好的信度和结构效度。以其中86名飞行员的安全绩效考核(安全规章、飞行作风、飞行技能和机组管理)数据和飞行员在每个飞行阶段的驾驶安全行为总体表现为效标变量, 以航线驾驶安全行为评价量表四个维度上的驾驶行为特征为预测变量分别进行层次回归分析, 结果表明: 航线驾驶安全行为多维评价量表具有良好的预测效度; 四个维度变量对维护不同飞行阶段的驾驶安全行为分别具有显著的贡献性, 且这种贡献性因飞行阶段的任务不同而有所差异。航线安全驾驶行为多维评价量表的构建, 不仅为航线飞行员安全驾驶行为规范性评估、机组驾驶行为问题诊断以及安全绩效考核提供了一套客观有效的评价工具, 而且也为有针对性的设计航线飞行员选拔模式和有效实施机组人员非技术性技能/CRM训练提供了一个良好的理论基础。

Developing a Multidimensional Scale to Assess Safety Behaviors in Airline Flight

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.