Date of Award

Summer 2023

Access Type

Thesis - Open Access

Degree Name

Master of Science in Aviation


College of Aviation

Committee Chair

Jennifer E. Thropp, Ph.D.

First Committee Member

Scott R. Winter, Ph.D.

College Dean

Alan J. Stolzer, Ph.D.


Risk-taking, a persistent topic of interest and concern in aviation, has been linked with unsafe behaviors and accidents. However, risk-taking propensity is a complex construct that encompasses numerous factors still being researched. Even within the limited research available about the factors affecting pilots’ risk-taking propensity, studies have yielded inconsistent results. Therefore, this quantitative study explores existing and novel factors that predict the propensity for risk-taking among general aviation (GA) pilots in the United States.

This study, conducted in two stages, involved developing a prediction model using backward stepwise regression to predict pilots’ risk propensity, followed by model fit testing using additional sampling to validate the predicted model. Data was gathered using surveys from multiple local Experimental Aircraft Association (EAA) chapters in Central Florida and from Embry-Riddle Aeronautical University, Daytona Beach campus. In Stage 1, the model was constructed based on data obtained from 100 participants. Stage 2 involved validating the model using responses from another 100 participants who answered the same set of questions as in Stage 1. Model validation encompassed three methods: correlation analysis, t-test, and cross-validity coefficient. The results from these analyses demonstrated a strong fit between the regression model and the Stage 2 data, affirming the accuracy of the prediction model.

The analysis identified a model comprising seven significant predictors among a set of 12, accounting for 76% of the variance, with an adjusted R2 of 75%, influencing the risk-taking propensity among GA pilots. These predictors included age, total flight hours, number of flight ratings, number of hazardous events, self-efficacy, psychological distress, and locus of control. Model prediction and cross-validation were employed to enhance the findings’ rigor and generalizability. Practical applications and suggested areas for future studies are also discussed.