Dollars or Cents: Impacts of Rescaling Data on a Mixed Logit Model with Normally and Lognormally Distributed Coefficients

It is a common practice to rescale data to assist the model estimation process. This paper describes a case where care is called upon when interpreting the results after rescaling. The case is shown as associated with the lognormally distributed coefficient in a mixed logit model. Implications of rescaling data on a normally distributed coefficient are also given for comparison.


Issue Date:
2004
Publication Type:
Working or Discussion Paper
Record Identifier:
http://ageconsearch.umn.edu/record/24108
PURL Identifier:
http://purl.umn.edu/24108
Total Pages:
12
Series Statement:
Staff Paper 04-01




 Record created 2017-04-01, last modified 2018-01-22

Fulltext:
Download fulltext
PDF

Rate this document:

Rate this document:
1
2
3
 
(Not yet reviewed)