How Effective Are Marker Variables at Predicting Attitudinal Factor Scores? An Out-of-Sample Evaluation

Term Start:

October 1, 2023

Term End:

May 31, 2024

Budget:

$93,544

Keywords:

Attitudes, Machine Learning, Marker Variables

Thrust Area(s):

Data Modeling and Analytic Tools

University Lead:

Georgia Institute of Technology

Researcher(s):

Patricia Mokhtarian

Despite the fact that our existing models are not up to the job of predicting travel behavior in today’s rapidly changing landscape, and despite considerable evidence that attitudes help us explain behavior more completely and more meaningfully, attitudes are nowhere to be found in practice-oriented travel demand forecasting models.  Two main objections have been raised to their inclusion:  they are too cumbersome to measure, and difficult-if-not-impossible to forecast.  This project would continue a line of research that focuses on overcoming the first objection.  Specifically, the plan is to use machine learning methods to train a prediction function on one survey dataset (the “donor sample”, and then apply that function to impute attitudes into another dataset (the “recipient sample”).  This keeps the recipient survey less burdensome on the respondent, while allowing the dataset to receive attitudinal information that would otherwise be absent.

Scroll to Top