Statalist The Stata Listserver

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

Re: st: probit- dprobit

From   Maarten buis <>
Subject   Re: st: probit- dprobit
Date   Fri, 11 May 2007 20:37:03 +0100 (BST)

> At 09:55 AM 5/11/2007, Carola Weiss wrote:
> >Iīm using the probit command to figure out the 
> >impact of tenure to work satisfaction (binary 
> >variable: 1= high work satisfaction).
> >I should estimate the marginal effects by using dprobit.
> >Can anybody tell me whatīs he differnece? I 
> >should only interpret  significance and 
> >direction of the effects. Therefore I only need the probit results,
> >donīt I?

--- Richard Williams <> wrote: 
> Why do you think you should only interpret 
> significance and direction of effects?  One of my 
> pet peeves is that people do indeed often only 
> look at sign and significance, and the 
> practical/substantive significance remains 
> unclear.  Marginal effects are one way of getting at substantive
> significance.

I totally agree with Richard. Significant is not the same as relevant.
If I am in a bad mood I tell whomever comes in my office for help that
if they get a non-significant result that only means that they haven't
collected enough data, and if they get a significant result that they
collected too much data. Don't get me wrong, significance testing is a
very elegant and smart solution to an impossible problem, but too often
people attach magical properties to p-values. Just looking at some
predicted probabilities, marginal effects, or odds ratios (I am
actually starting to appreciate those, though I admit they are an
acquired taste)  can tell you so much more. And don't forget graphs!!

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

Yahoo! Answers - Got a question? Someone out there knows the answer. Try it
*   For searches and help try:

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