Hello colleagues,
I would like to ask you what will be an appropriate econometric model for predicting an event that has never occurred before. For example the election of an extreme political party, which has never before been in government. To be clear we have no data for the dependent variable, but several independent variables. The objective of the study is to predict the probability of an event occurring.
Thank you and kind regards,
George K. Zestos
I agree with Florian Kaulich - I don't see how it could be done with a categorical measurement. Also in for instance multinomial logit or conditional logit models - that would allow you to test the possibility of a certain outcome to occur (conditional on values of the other options) - you would need at least some 'successes'.
So yes using a vote share or for instance seats in parliament, you could model the obtained results, and use this to predict in which circumstances a certain party would obtain enough votes to win the election (or enter Parliament, or what you're interested in).
- Badges
- Science topic
- Similar topics
- Psychology