[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: help with fractional logit and marginal effects

From   Maarten buis <>
Subject   Re: st: help with fractional logit and marginal effects
Date   Tue, 15 Dec 2009 13:35:25 -0800 (PST)

--- On Tue, 15/12/09, george joseph wrote:
> The dependent variable, pengscore1 originally ranges from
> 0-100(percentage of correct answers) which I rescaled to 0-1.
> neavar are independent variables denoting school characteristics like
> teacher pupil ratio, penetage of trained teachers , urban etc. My
> question is the folllowing.
> 1. Since I rescaled the dependent variable, how do I get
> the correct marginal effects- should I use- mfx compute or mfx compute
> , eyex. Since the dependent variable is rescaled how do I interpret
> the marginal effects.

It is all much simpler than you think, the transformation from percentages
to proportions is just a linear transformation (divide by 100), so that
does not change a thing. All you need to take care of that the unit of the
dependent variable is now proportion and not percentage. If you want you
can just multiply everything (marginal effects, their standard errors and
confidence interval limits) with 100, to change the unit to percentages 

*------------------------ begin example ------------------------------
use "";, clear
glm governing minorityleft noleft popdens houseval, link(logit) robust
mfx, at(minorityleft=0 noleft=0)
*------------------------ end example --------------------------------

Hope this helps,

Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen


*   For searches and help try:

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