Estimating a model of airline passenger choice using grouped booking data and least squares regression

In this article, we estimate a model of airline passenger choice using grouped booking data and least squares regression rather than the standard approach based on individual PNR‐level booking data and maximum likelihood estimation. We adapt the little‐known and somewhat forgotten transformation techniques proposed by Berkson and Theil while correcting for heteroskedasticity using weighted least squares. We illustrate how to apply the methodology for revenue management applications for data grouped by path and fare class at several stages of the booking process. This approach has the practical advantages of relying on data at the level of aggregation used in most revenue management systems, reducing the size of the dataset and simplifying the estimation process.