Effect of previous manual flying on manual handling performance.
Ebbatson et al. (2010) completed research into the relationship between manual handling performance of aircraft and recent flying experience (Ebbatson, Harris, Huddlestone & Sears, 2010). The research concluded that pilots who had flown more sectors in the week prior to the testing had a smaller mean heading error than those pilots who had flown fewer sectors in relation to the straight and level flight segment testing (p = 0.002). The below article provides an overview and summary of their results. A main hypothesis was presented that pilots previous manual flying sectors would have no effect on their manual handling performance, with an alternate hypothesis presented that previous manual flying sectors would increase manual handling performance.
The sample utilised in this research were all professional airline pilots with a type endorsement for a Boeing 737-300/400/500 aircraft from one participating airline.
A table was created showing the results for flight data-derived performance measurements separated into the flight phases of the exercise. The sample size is 49 as a result of data capture failures during the testing.
The areas tested by Ebbatson et al. (2010) in their research were as follows:
Many of the tested values have little correlation to each other, however the authors of the article state that a significant negative correlation exists for the mean error in heading (r = -0.413), and the frequency of rudder inputs (r = -0.443). It is also stated that the frequencies of yaw control inputs were related to the number of sectors flown prior to testing (r = -0.380). These correlations are deemed significant due to the size of the sample, however the probability value of these parameters are, p = 0.002, p = 0.001 and p = 0.004 respectively. Both the probability value and negative correlations tend to show an acceptance of the alternate hypothesis that pilots who had flown more sectors previously showed smaller mean averages to those that had not.
Both the Control Wheel and Control Column results displayed no significant relationship in correlation.
The testing was completed with the assistance of a type rated examiner in a Boeing 737-300 full flight simulator equipped with six degrees of freedom hydraulic motion cueing system. The simulator was approved to JAR STD 1A level D. The test was completed once for each pilot at the conclusion of their annual licence proficiency check and used frequency based measures.
The sample consisted of 32 first officers, 27 captains and seven training captains. The mean age of the sample was 40 years, SD 9 years, and a mean flying hours of 5887 hours, SD 3839 hours. For the above figures, the mean experience flying highly automated aircraft was 3597 hours, SD 2803 hours. The mean experience of the sample flying Boeing 737-300/400/500 aircraft was 2269 hours, SD 1772 hours. During the week prior to the research being conducted, the scope of the sectors flown was between 0 and 16 sectors, with a mean number of sectors of 5.45, SD 3.92 sectors.
Weather information reported to the pilots was marginal, allowing them to proceed with the approach but making the outcome at 'decision height' unpredictable as to whether the pilot would perform a go-around or not. The simulated cloud base was set slightly lower than reported, and subsequently was also below the 'decision height' for pilots, requiring the crew to perform a missed approach procedure due to lack of visibility of the runway.
The order in which the crew members were assigned pilot flying duties were also varied over the testing.
While three out of the five parameters used to show correlation between pilots previous manual flying do show a significant correlation based on the sample size of the research, it is difficult to draw a satisfactory conclusion from the research done by Ebbatson et al. (2010). The authors have attempted to infer the alternate hypothesis that previous manual flying experience led to greater manual flying performance, but two of the five parameters can also infer the main hypothesis that sectors previously manually flown have little to no bearing on manual handling performance.