[d] panel · [s] shuffle · [f] fullscreen · [r] reverse · [m] mode

Model Actions,
Not Words

Models listen to what people say. We predict what people do. Bridging the psychology of words and the mathematics of action.

Explore the framework↗︎

// The truth gap

Words vs. Decision

Standard economic models assume we are rational robots. We aren’t. We have biases, moods, and memories.

Hybrid Choice Models use the rigor of economics to contain the messiness of psychology, transforming vague feelings into precise market predictions.

Subconscious models the DO, not the SAY.

Standard economic models assume rational agents. Hybrid Choice Models bridge psychology and economics.

Quantifying the unobservable

The HCM framework

Explanatory variables feed into latent classes and latent variables. Latent variables are measured through indicators - survey responses that signal hidden traits like risk aversion or brand loyalty.

The decision process combines revealed preference (what people do) with stated preference (what they say) through utility maximization.

Based on Ben-Akiva & McFadden’s econometric framework for modeling choice under uncertainty.

Research domains

01 ::

Behavioral choice analysis

Psychology: explains the ‘why’ but lacks a mathematical framework for forecasting.

02 ::

Predictive choice models

Economics: tracks the ‘what’ using revealed market data, often ignoring the human element.

03 ::

Random Utility Model

The mathematical core. Assumes rational utility maximization - powerful, but blind to psychological nuance.

04 ::

Hybrid Choice Models

The bridge. Quantifies latent psychological variables within an econometric framework to predict real-world decisions.

Model components

01 ::

Explanatory variables

Observable data - age, income, price - that feeds into the model as input.

02 ::

Latent classes

Hidden segments like ‘trendsetters’ or ‘value seekers’ discovered through behavioral clustering.

03 ::

Latent variables

Hidden psychological traits - risk aversion, loyalty, eco consciousness - measured through survey indicators.

04 ::

Stated preference

Survey data: what people say they will do. Descriptive but often unreliable for prediction.

05 ::

Revealed preference

Market data: what people actually do. The ground truth for behavioral modeling.

06 ::

Decision process

Utility maximization: the mathematical engine that combines all inputs to predict choice probabilities.