[Limdep Nlogit List] Logit with repeated measures

[Andrea Vocino]

Sorry for the basic question. I am analysing a discrete choice experiment where respondents were asked to make choices between two different alternatives. The experiment design entailed a randomised block where each respondent was asked to make nine choices (e.g., repeated measures). As a result, even though the blocks were randomised, assuming the observations being IID might induce to create bias estimates as there might be the need for accounting for certain correlation structures among the clustered observations and for overdispersion among discrete responses modelled with binomial models (Agresti 1990; Hartzel et al. 2001). I am a little confused about what logit method to use as the applied literature has used a little bit of everything. The alternatives that come up to my mind are



1)  Conditional (fixed effects) logit model

2)  Random effects logit model, although the econometric literature suggests to perform the Hausman test to determine whether the fixed or random effects model is more appropriate

3)  A "normal" logit with a sandwich estimator for clustered data (where the "respondent" is the "cluster")

4)  Generalised estimating equations (GEEs) with the possibility of using different (e.g. independent, exchangeable etc) correlation matrices.

5)  I am sure there must be another dozen that could be used  . . .



Thanks in advance for your time



Andrea
_______________________________________________________________________________

Andrea Vocino: email<mailto:andrea.vocino at deakin.edu.au> | web<http://andrea.vocino.name/> | linkedin<http://au.linkedin.com/pub/andrea-vocino/B/B11/A3B> | facebook<http://www.facebook.com/andrea.vocino> | p +61 3 925 17408 | f +61 3 925 17083 | m 041 113 6087
Faculty of Business & L aw<http://www.deakin.edu.au/buslaw/> | School of Management & Marketing<http://www.deakin.edu.au/buslaw/management-marketing/>
Deakin University CRICOS Provider Code 00113B

Important Notice: The contents of this email are intended solely for the named addressee and are confidential and
may be subject to legal privilege; any unauthorised use, reproduction or storage of the contents is expressly prohibited.
If you have received this email in error, please delete it and any attachments immediately and advise the sender by
return email or telephone. Deakin University does not warrant that this email and any attachments are error or virus
free