Statalist


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

Re: st: mfx


From   "Chiara Mussida" <cmussida@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: mfx
Date   Mon, 21 Jul 2008 15:40:56 +0200

Sure, it is a plot for a continuous variable keeping constant at its
mean another continuous variable. It is what I asked and you kindly
suggested to me: thanks for this! But now, for another estimation I
would be interested in plotting the relation between my binary
dependent variable and one continuous covariate from the outset: to
see how an infinitesimal change in the continuous covariate affects
the probability of making a specific transition (which is my dep
variable). Is it impossible?
Thanks a lot

2008/7/21 Arne Risa Hole <arnehole@gmail.com>:
> Chiara,
>
> Try to be a bit more specific when explaining what you want to do...
> initially you asked for a plot of the marginal effect of a contionous
> variable which is what the example I sent earlier does. With a
> discrete 0-1 dummy there is nothing to plot as there is only one
> possible change from 0-1.
>
> Arne >
> 2008/7/21 Chiara Mussida <cmussida@gmail.com>:
>> At this stage it is better to ask:
>> the command below only allow to get the mfx for continuous variables?
>> this is also because, by trying with a discrete, you get the error
>> message below. Again: is there a way to obtain a range for y? it would
>> be nice, since the dep var in my model covers a wide range if values!!
>>
>> Thanks and Sorry for the absence of details below!
>>
>> 2008/7/21 Arne Risa Hole <arnehole@gmail.com>:
>>> Chiara,
>>>
>>> As has been pointed out many times before on the list, the Statalist
>>> FAQ section 3.3 says:
>>>
>>> Say exactly what you typed and exactly what Stata typed (or did) in
>>> response. N.B. exactly! If you can, reproduce the error with one of
>>> Stata's provided datasets or a simple concocted dataset that you
>>> include in your posting.?
>>>
>>> If you do that someone might be able to find the error.
>>>
>>> Arne
>>>
>>> 2008/7/21 Chiara Mussida <cmussida@gmail.com>:
>>>> Thanks for all these suggestions!!
>>>> the twoway function plot perfectly works with the probit formulation.
>>>> But when I try with a biprobit it says:
>>>> ; is not a twoway plot type
>>>> r(198);
>>>> without claryfing the reasons!
>>>>
>>>> 2008/7/19 Joao Ricardo F. Lima <jricardofl@gmail.com>:
>>>>> Arne
>>>>>
>>>>> thanks for your answer. Now I understand that x represents the mpg
>>>>> variable. I was curious because I had changed the code to:
>>>>>
>>>>> #delimit ;
>>>>> twoway (function y = _b[mpg]*normden(_b[price]*`mean_price'+
>>>>> _b[mpg]*mpg+_b[_cons]), range(mpg)) ;
>>>>> #delimit cr
>>>>>
>>>>> and the graph generated was different.
>>>>>
>>>>> Best regards,
>>>>>
>>>>> Joao Lima
>>>>>
>>>>> 2008/7/19 Arne Risa Hole <arnehole@gmail.com>:
>>>>>> Joao,
>>>>>>
>>>>>> What the code below does is plotting the marginal effect of the mpg
>>>>>> variable evaluated at different values, holding the price variable
>>>>>> constant at its mean. The x in the code represents the mpg variable -
>>>>>> see -help twoway function-.
>>>>>>
>>>>>> Perhaps this is clearer:
>>>>>>
>>>>>> sysuse auto
>>>>>> probit foreign price mpg
>>>>>> mfx
>>>>>> sum price
>>>>>> local mean_price = r(mean)
>>>>>> #delimit ;
>>>>>> twoway (function y = _b[mpg]*normden(_b[price]*`mean_price'+
>>>>>> _b[mpg]*x+_b[_cons]), range(mpg) ytitle(Marginal effect)
>>>>>> xtitle(Miles per gallon)) ;
>>>>>> #delimit cr
>>>>>>
>>>>>> Arne
>>>>>>
>>>>>> 2008/7/19 Joao Ricardo F. Lima <jricardofl@gmail.com>:
>>>>>>> Dear,
>>>>>>>
>>>>>>> sorry the simple question. I am curious why I have to use "x" instead
>>>>>>> mpg (mean) in the twoway graph (..._b[mpg]*x+_b...)? Or better, what
>>>>>>> Stata understand when I type x in this command?
>>>>>>>
>>>>>>> Thanks a lot,
>>>>>>>
>>>>>>> Joao Ricardo
>>>>>>>
>>>>>>> 2008/7/18 Arne Risa Hole <arnehole@gmail.com>:
>>>>>>>> Hi Chiara,
>>>>>>>>
>>>>>>>> How about this?
>>>>>>>>
>>>>>>>> sysuse auto
>>>>>>>> probit foreign price mpg
>>>>>>>> mfx
>>>>>>>> sum price
>>>>>>>> local mean_price = r(mean)
>>>>>>>> #delimit ;
>>>>>>>> twoway (function y = _b[mpg]*normden(_b[price]*`mean_price'+
>>>>>>>> _b[mpg]*x+_b[_cons]), range(mpg)) ;
>>>>>>>> #delimit cr
>>>>>>>>
>>>>>>>> This plots the marginal effect of the mpg variable over its range
>>>>>>>> following a probit regression. The logic is the same with -biprobit-
>>>>>>>> but the formula for the marginal effect will of course be a little
>>>>>>>> more intricate in that case.
>>>>>>>>
>>>>>>>> Arne
>>>>>>>>
>>>>>>>> 2008/7/18 Chiara Mussida <cmussida@gmail.com>:
>>>>>>>>> Dear All,nuous variab
>>>>>>>>> is there a way to "plot" the mfx of a continuous variable (after a
>>>>>>>>> biprobit) to infer its behaviour over the complete range of values and
>>>>>>>>> not only for the mean of the independent variables (which is the
>>>>>>>>> default)?
>>>>>>>>>
>>>>>>>>> Thanks
>>>>>>>>>
>>>>>>>>> --
>>>>>>>>> Chiara Mussida
>>>>>>>>> PhD candidate
>>>>>>>>> Doctoral school of Economic Policy
>>>>>>>>> Catholic University, piacenza (Italy)
>>>>>>>>> *
>>>>>>>>> *   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/
>>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> --
>>>>>>> -------------------------------
>>>>>>> Joao Ricardo Lima
>>>>>>> Professor
>>>>>>> UFPB-CCA-DCFS
>>>>>>> +553138923914
>>>>>>> -------------------------------
>>>>>>> *
>>>>>>> *   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/
>>>>>>
>>>>>
>>>>>
>>>>>
>>>>> --
>>>>> -------------------------------
>>>>> Joao Ricardo Lima
>>>>> Professor
>>>>> UFPB-CCA-DCFS
>>>>> +553138923914
>>>>> -------------------------------
>>>>> *
>>>>> *   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/
>>>>>
>>>>
>>>>
>>>>
>>>> --
>>>> Chiara Mussida
>>>> PhD candidate
>>>> Doctoral school of Economic Policy
>>>> Catholic University, piacenza (Italy)
>>>> *
>>>> *   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/
>>>
>>
>>
>>
>> --
>> Chiara Mussida
>> PhD candidate
>> Doctoral school of Economic Policy
>> Catholic University, piacenza (Italy)
>> *
>> *   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/
>



-- 
Chiara Mussida
PhD candidate
Doctoral school of Economic Policy
Catholic University, piacenza (Italy)
*
*   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 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index