Commercial air travel is regarded as the safest mode of transportation known to humankind; however, every year people lose their lives from aircraft accidents and incidents. In addition, the financial impact of an air disaster can destroy an airline organisation. Studies have found that in adverse events involving highly advanced aircraft employing complex automation, human factor issues, and particularly pilot training, continue to play a significant causal role. Special attention should therefore be paid to the training of airline pilots, who are ultimately the last line of defence in aircraft operations. Airline pilots’ perceptions of the training climate associated with advanced aircraft can be a pervasive and powerful determinant of training outcomes and eventual flight deck behaviour. The study undertook to develop a valid and reliable instrument to measure airline pilots’ perceptions of the training climate associated with advanced aircraft equipped with highly complex automation. The goal was to construct a questionnaire by operationalizing an unobserved hypothesised construct (perceptions of the advanced automated aircraft training climate) based on three levels of analysis (the microsphere, mesosphere and macrosphere). The study also attempted to explore the statistical relationship between the demographic variables of the respondents and the latent factors of the construct. In order to meet the research objectives, the study began with a thorough review of the current literature on the topic to develop a systems model of the main construct under investigation. The review included a critique of the theory on organisational climate, learning, training and education, of historical data on aircraft automation, of human factors, and of aircraft accident investigation principles and case studies. The objectives of the research were fulfilled by strictly observing a positivist paradigm, and engaging in a quantitative exploration, triangulating methods with data captured from a purposive sample of the target population. The empirical study was completed in four phases. Firstly, the research construct was operationalized and the items in the proposed questionnaire validated by a panel of subject matter experts using Lawshe’s (1975) content validity ratio (CVR) technique. Inter-rater bias was assessed using Cochran’s Q test. This application resulted in the retention of 42 items. Secondly, factor analysis and item analysis was performed on the responses of the respondents for the development of the final 33 item measurement instrument. Thirdly, to explore the relationship between the demographic variables and the latent factors of the main construct, an appropriate non-parametric family of statistics was selected to gain a deeper understanding of the phenomena associated with the data. Finally, a logistic regression analysis that included specific demographic variables was performed for the development of a model to predict a pilot’s perception of the training climate associated with advanced automated aircraft. A non-probability purposive sample of 17 subject matter experts and 229 qualified South African airline pilots was used to accomplish the goals of the study. The underlying structure of the advanced Automated Aircraft Training Climate Questionnaire (AATC-Q) was derived from the results of a Principal Axis Factor (PAF) analysis using a promax (Kappa-4) rotation. The number of factors extracted from the data set was based on a modified version of Horn’s (1965) parallel analysis, namely the Monte Carlo simulation algorithm designed by O’Connor (2000). Three core factors explained most of the underlying variability in the main construct. The first factor was a composite at the macro and meso levels of analysis, whilst the second and third factors became fragmented at the micro level of analysis. These three factors were then labelled Organisational Professionalism, Intrinsic Motivation and Individual Control of Training Outcomes. The quality and rigour of the derived scale were demonstrated by its content and construct validity. Overall, satisfactory results from computing Cronbach’s coefficient alpha showed that the measurement scale was also reliable. The effect of the demographic variables on airline pilots’ perceptions of the advanced automated aircraft training climate was determined by computing relationships and comparing the responses from different categorised subsets with one another, by means of a non-parametric MANOVA and non-parametric analysis of variance. The results of these tests revealed that Flight Deck Position, Size of the Airline, Computer Literacy and Flight Experience had a significant effect on a pilot’s perception of the training climate. Results from a logistic regression model indicated that the interaction between pilots’ experiences and their perceived level of computer literacy (on a sigmoid curve), their actual experience in advanced aircraft, and their preferences for route and simulator training, were related to whether a pilot perceived the advanced aircraft training climate as favourable or not. The overall percentage of cases for which the dependent variable was correctly predicted by the regression model was computed at 63.8%. This study represents a vital step toward an understanding of the dimensionality of the learning, education and training for, and the actual operation of, highly advanced commercial aircraft, which employ complex automation. The results provide sufficient empirical evidence to suggest that the research findings may be of particular interest to aviation psychologists, aviation safety practitioners, and airlines engaged in training pilots to operate advanced aircraft. / Thesis (PhD)--University of Pretoria, 2012. / Human Resource Management / unrestricted

abstract: From 2001-2011, the General Aviation (GA) fatal accident rate remained unchanged (Duquette & Dorr, 2014) with an overall stagnant accident rate between 2004 and 2013. The leading cause, loss of control in flight (NTSB, 2015b & 2015c) due to pilot inability to recognize approach to stall/spin conditions (NTSB, 2015b & 2016b). In 2013, there were 1,224 GA accidents in the U.S., accounting for 94% of all U.S. aviation accidents and 90% of all U.S. aviation fatalities that year (NTSB, 2015c). Aviation entails multiple challenges for pilots related to task management, procedural errors, perceptual distortions, and cognitive discrepancies. While machine errors in airplanes have continued to decrease over the years, human error still has not (NTSB, 2013). A preliminary analysis of a PC-based, Garmin G1000 flight deck was conducted with 3 professional pilots. Analyses revealed increased task load, opportunities for distraction, confusing perceptual ques, and hindered cognitive performance. Complex usage problems were deeply ingrained in the functionality of the system, forcing pilots to use fallible work arounds, add unnecessary steps, and memorize knob turns or button pushes. Modern computing now has the potential to free GA cockpit designs from knobs, soft keys, or limited display options. Dynamic digital displays might include changes in instrumentation or menu structuring depending on the phase of flight. Airspeed indicators could increase in size to become more salient during landing, simultaneously highlighting pitch angle on Attitude Indicators and automatically decluttering unnecessary information for landing. Likewise, Angle-of-Attack indicators demonstrate a great safety and performance advantage for pilots (Duquette & Dorr, 2014; NTSB, 2015b & 2016b), an instrument typically found in military platforms and now the Icon A5, light-sport aircraft (Icon, 2016). How does the design of pilots’ environment—the cockpit—further influence their efficiency and effectiveness? To explore the possibilities for small aircraft displays, a participatory design investigation was conducted with 9 qualified instrument pilots. Aviators designed mock cockpits on a PC using pictorial cutouts of analog (e.g., mechanical dials) and digital (e.g., dynamic displays) controls. Data was analyzed qualitatively and compared to similar work. Finally, a template for GA displays was developed based on pilot input. / Dissertation/Thesis / Masters Thesis Applied Psychology 2016