Risk Perception Comparison Between Experts and Novices.
In 2004, Thomson, Onkal, Avcioglu & Goodwin conducted research into the risk perception differences between, expert and novice pilots in relation to helicopter operations. The study reports possible perceptual differences between the two different participant groups. The risk perception comparisons in this study were made through the administration of questionnaires, where each participant was required to rate their perceived risk of certain incidents and scenarios.
Illustration 1 displays the mean ratings of both participant groups. The mean results displayed in Illustration 1 show that for the majority of the incidents listed, the novice participant group rated the perceived risks of the incidents higher than that of the expert participant group.
Illustration 1: Mean ratings from participant groups.
Illustration 2 provides a clearer picture of this trend by comparing the average ratings over the entire questionnaire of both participant groups. The average ratings displayed in illustration 2 show that on average the novice participant group perceived the incidents to be 6.17% riskier than the expert pilots.
|Illustration 2: Average Participant Group Ratings for Questionnaire.|
|Expert Pilots||Novice Pilots|
The average values for the standard deviations of both participant groups, as displayed in illustration 3 are relatively large. This indicates that the data also has a large range which in turn suggests that the data is less reliable. A basic statistics rule claims that: a hypothesis should not supported if both standard deviations are larger than the difference between the two means. In this case the difference between the means, as previously mentioned is much lower than that of the standard deviations (6.17< 25.46 & 6.17< 23.69). Consequently, this indicated that the data does not support the hypothesis. In addition the p-value shown in illustration 3 is calculated to be 0.161, which also concludes that there is no evidence that the null hypothesis does not hold.
|Illustration 3: Standard Deviation Values and p-Values From Results|
|Incident||Expert S.D.||Novice S.D.||p-Value|
|Tail rotor shaft breakage||28.0||27.5||0.178|
|Main rotor blade failures||24.5||30.7||0.012|
|Transmission system failures||25.0||17.3||0.332|
|Hydraulic system failures||18.1||19.3||0.014|
|Electric system failures||22.8||20.2||0.000|
|Wire/water/ground contact at NOE||25.3||22.3||0.432|
|Loss of tail rotor thrust||24.1||24||0.323|
|Accidents due to excessive usage in hover & taxi||25.9||21.1||0.102|
The samples used very specific parameters, which would have restricted any significant findings to a very specific population. However, the study gives direction for future research as the authors recognize that more research is need before generalizations and conclusions can be drawn from the data; in addition the authors identify important areas for future research.
Exploratory research into the risk perception differences between expert and novice helicopter pilots.
The study was comprised of two independent non-probability samples. The first sample (expert pilots) consisted of 36 experienced helicopter pilots, with a mean age of 34 years. The second sample (novice pilots) consisted of 28 pilots, with a mean age of 23. All participants in the study were Turkish males who were or had previously attended the Turkish Military Academy and Army Aviation School.
Thomson, Onkal, Avcioglu & Goodwin (2004), administered a questionnaire containing 13 incidents, which were to be rated on a scale of 0 (no risk) to 100 (maximum risk).
Two of the study's additional items were omitted for analysis from this article, these include: Aviation Risk Characteristics as it was open to multiple interpretations and lacked the ability to make effective comparisons between the two participant groups; additionally, Hypothetical Scenarios due to the inconsequential data relative to the study.
The original article provided only individual mean ratings for each incident list in the questionnaire. Though the use of average rating scores from the entire questionnaire, the reader can effectively make comparison between the two participant groups.
Want to know more?
Read the full article that was published in the: Risk Analysis, Volume 24, Issue 6, pages 1585–1595, December 2004
Kinney, Lucas J. (2013).