Hybrid choice models

Hybrid choice models (HCMs) are extensions of classical discrete-choice models that combine what analysts can observe—prices, incomes and travel times—with what they cannot measure directly, such as attitudes, perceptions and habits. Unlike standard models, which recover preferences only from observed choices, an HCM links those choices to one or more latent variables—statistical constructs that represent psychological traits. Integrating such constructs into a random-utility framework improves both behavioural realism and predictive accuracy.