16 Qualitative Choice Models

The last chapter introduced binary choice models that answer questions such as:

  • Did you vote during the last election?
  • Does an individual recidivate after being released from prison?

In this chapter, cases of dependent variables with more than two categories are considered. The categorical dependent variables can be with or without natural ordering. Here are some examples of natural ordering:

  • Level of happiness: Very happy, happy, okay, or sad
  • Intent to get vaccinated: Definitely yes, probably yes, probably no, definitely no

Examples without natural ordering:

  • Type of car to purchase: Passenger car, Pick-up truck, van, convertible

Those models are also known as Discrete Choice Models. The packages required for this section are mlogit, erer, MASS, AER, glmmML, and nnet. The data required are fpdata and evdata in addition to Fishing and TravelMode which are part of the packages mlogit and AER. You can load them into R with the following commands:

  • data("Fishing",package="mlogit")
  • data("TravelMode",package="AER")

There is also a YouTube Video associated with this chapter.