On Least Squares Estimation When the Dependent Variable is Grouped
Models estimated from censored samples are now familar in the econometrics literature. For many cases Least Squares approximations to the Maximum Likelihood estimators are now well established. This paper is concerned with a more general problem ; that of estimating an equation on the basis of data in which the dependent variably is only observed to fall in a certain range on a continuous scale, its actual value remaining unobserved. The date are also censored in the usual sense in that both end ranges are assumed to be open-ended. A number of Least Square approximations to the Maximum Likelihood estimator are derived and compared. The results of Greene (1981) on the asymptotic bias of OLS are extended to this case. The question of information loss as a result of the grouping is also considered.
(This abstract was borrowed from another version of this item.)
|Date of creation:||Nov 1982|
|Date of revision:|
|Contact details of provider:|| Postal: |
Phone: 609 258-4041
Fax: 609 258-2907
Web page: http://www.irs.princeton.edu/
More information through EDIRC
When requesting a correction, please mention this item's handle: RePEc:pri:indrel:dsp01j9602061c. See general information about how to correct material in RePEc.
For technical questions regarding this item, or to correct its authors, title, abstract, bibliographic or download information, contact: (David Long)
If references are entirely missing, you can add them using this form.