Passenger security screening is a continuously evolving process and not much is known about how passengers perceive its effectiveness in keeping air travel safe. The provided thesis consists of two manuscripts that expand the literature on passenger perception of safety through the application of advanced econometric techniques. This thesis aims to determine the factors that influence passenger level of confidence in the abilities of the TSA to keep air travel safe and identify how these might differ between gender. An extension of the mixed logit model is applied to better estimate the data and capture unobserved heterogeneity (random parameters) as well as account for possible heterogeneity in the means.
The first manuscript explores this concept of passenger perception of safety using the provided Omnibus Household Survey data. Current literature is lacking regarding how passengers perceive safety in their air travel experiences. This understanding is integral to create a more comfortable environment for all passengers. A mixed logit model is estimated for the data which found various factors to influence a passenger’s level of confidence in the TSA’s abilities. This was followed by the incorporation of an extension of the mixed logit model accounting for the possibility of heterogeneity to be present in the means of the random parameters. A likelihood ratio test concluded that the more advanced mixed logit model that captured heterogeneity in the means best fits the data. This manuscript creates a base of knowledge for modeling passenger perception of safety which is further explored in the following manuscript using similar advanced methodological techniques.
The second manuscript expands upon the findings from the first manuscript to analyze if and how gender plays a role in the factors that influence passenger perception of safety. It is important to identify this difference in order to best accommodate for all air travel passengers. To begin, a mixed logit model is estimated for the female and male participants separately. A few variables were found to be significant in both the female and male models, but the majority of variables were not present in both gender models. Next, an extension of the model was estimated that allowed for heterogeneity in the means of the random parameters. The female model was found to have heterogeneity in the means present while the male model resulted in none. A variety of calculated likelihood ratio tests indicated the separation of male and female to be valid and significant variables to not be transferable between genders. Furthermore, it was concluded that the mixed logit model with heterogeneity in the means best fit the female data.
This thesis used an advanced methodology to identify factors that influence passenger level of confidence in air travel safety and furthermore, concluded that gender should be modeled separately. These findings are integral in understanding how passengers perceive the effectiveness of passenger security screening in order to improve the process. Ideally, the results from these manuscripts can be shared with airport management and the TSA to create a more comfortable experience and accommodating environment for all air travel passengers.