No calibration required

Date of publication: September 19, 2004


Using data from North America and Western Europe, the authors demonstrate the further usefulness of discrete choice models in that they do not yield substantively different results based on the medium the respondent uses (on-line or off-line). This characteristic provides an opportunity for forecasting models based on discrete choice probabilities to gain the upper hand on normative models that require extensive calibration of on-line data in order to compare to the off-line equivalent.

  • PDF
  • This could also be of interest