Statalist


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

Re: st: interpretation of interaction terms


From   René Geppert <[email protected]>
To   [email protected]
Subject   Re: st: interpretation of interaction terms
Date   Mon, 15 Dec 2008 16:12:29 +0100

In an all linear specification, regression coefficients are usually interpreted as marginal effects. Accordingly, in a log-log-specification, regression coefficients can be thought of elasticities. And for a lin-log-model, hence, the coefficient represent something like semi-elasticities.

-René

David Airey schrieb:
Yes, but I think the basic understanding or interpretation of regression coefficients for continuous or binary variables is not altered when a variable has been log transformed. It is true that interactions can be present or absent depending on whether or not a variable has been transformed, because interactions are scale dependent.

-Dave

On Dec 14, 2008, at 7:31 PM, Shehzad Ali wrote:

Hi,

Could someone please help me interpret the following:

1. Coefficient on an interaction term for a dummy and log-transformed variable in an OLS model when the dependent variable is in natural units or is also log-transformed

2. Coefficient on an interaction term for an untransformed continuous and log-transformed variable in an OLS model when the dependent variable is in natural units or is also log-transformed

3. Coefficient on an interaction term between a dummy and log-transformed variable in a probit regression

Thank you,

Shehzad
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index