[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: Logit/probit across different groups

From   Maarten buis <>
Subject   Re: st: Logit/probit across different groups
Date   Thu, 22 Nov 2007 15:33:45 +0000 (GMT)

--- tom albrecht <> wrote:
> The observations in my data are comparable to job applications. There
> are several different jobs and one group of applications for each
> job. From these competing applications one is picked for the job.
> Now I want to determine, which characteristics of an application have
> a significant influence on the probability that the application is
> getting picked - across all jobs.
> So I considered a logit or probit model with robust standard errors
> clustering on job level as well as including random-effects or
> fixed-effects for the jobs.
> My problem is: how do I correct for the differences of the
> characteristics between jobs?
> Let's say one characteristic is that people ask for a desired income
> in their application. So I have jobs where applicants ask for an
> income of 1000$ on average and groups where they ask for $ 1 million.
> Could I just calculate the income asked for relative to the group
> average or should apply a z-transformation? Are there standard
> procedures for this case or more advanced models?

If I understant correctly you have several person choose between
several jobs, and you want to control for characteristics of the person
and of the job. This sounds to me like a problem for -clogit-.

Hope this helps,

Maarten L. Buis
Department of Social Research Methodology
Vrije Universiteit Amsterdam
Boelelaan 1081
1081 HV Amsterdam
The Netherlands

visiting address:
Buitenveldertselaan 3 (Metropolitan), room Z434

+31 20 5986715

Want ideas for reducing your carbon footprint? Visit Yahoo! For Good
*   For searches and help try:

© Copyright 1996–2017 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index