Statalist


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

Re: st: How to test for difference in predicted prob after prvalue


From   "Kristin J. Kleinjans" <kkleinjans@econ.dk>
To   <statalist@hsphsun2.harvard.edu>
Subject   Re: st: How to test for difference in predicted prob after prvalue
Date   Sun, 9 Dec 2007 02:12:11 +0100 (CET)

Dear Sebastian and Richard,
this works well - thanks a lot for your help.

Best wishes,
Kristin


> Hi Kristin,
>
> as Richard pointed out: sounds like you are interested in the discrete
> change effect(s) in the probabilities for "sex" (or whatever the name of
>  your variable is, e.g. "male" to stick with Richard's example).
>
> You might want to look (again) at page 249 in Long/Freese (2006 [2nd
> edition], section 6.6.3). They describe exactly what Richard showed in
> his example just for race instead of sex (using options -save- and
> -diff- to calculate the discrete change effect).
>
> Note the option -rest()- which allows you to specify different values
> for your independent variables. Note further that you will end up with
> different results (different magnitude) for the discrete change when
> calculating it at different values of your covariates.
>
> Say hi to good old Europe for me,
> Sebastian (@IU Bloomington)
>
> Richard Williams schrieb:
>> At 09:55 PM 12/6/2007, Kristin J. Kleinjans wrote:
>>> Dear Statalist,
>>> I run 'mlogit' and use 'prvalue' to get predicted probabilities of
>>> outcomes for individuals with different characteristics (as described
>>> in Long and Freeses book on regressions for categorical variables).
>>> 'prvalue'
>>> also returns the confidence intervals of the predicted probabilities.
>>> I would like to test whether the predicted probabilities of a
>>> specific outcome are different for, let's say, women with the
>>> characteristics of men and average women. I hope somebody can help me
>>> on how to do this?
>>>
>>> Thanks a lot and best regards,
>>> Kristin
>>
>> See the help for -prvalue-. The -diff- option might give you what you
>> want.  Look to see whether 0 falls within the confidence interval for
>> the change; if not, the differences are statistically significant.
>> Example:
>>
>> . use "http://www.indiana.edu/~jslsoc/stata/spex_data/ordwarm2.dta";
>> (77 & 89 General Social Survey)
>>
>> . quietly mlogit  warm yr89 male white age ed prst
>>
>> . quietly prvalue, x(male=0) delta save
>>
>> . prvalue, x(male=1) delta diff
>>
>> mlogit: Change in Predictions for warm
>>
>> Confidence intervals by delta method
>>
>>                      Current     Saved    Change   95% CI for Change
>>   Pr(y=1SD|x):        0.1311    0.0962    0.0350  [ 0.0094,   0.0606]
>> Pr(y=2D|x):         0.3952    0.2621    0.1331  [ 0.0935,   0.1726]
>> Pr(y=4SA|x):        0.0994    0.2484   -0.1491  [-0.1795,  -0.1186]
>> Pr(y=3A|x):         0.3743    0.3933   -0.0190  [-0.0601,   0.0221]
>>
>>                yr89       male      white        age         ed
>> prst
>> Current=  .39860445          1   .8765809  44.935456  12.218055
>> 39.585259
>>   Saved=  .39860445          0   .8765809  44.935456  12.218055
>> 39.585259
>>    Diff=          0          1          0          0
>> 0          0
>>
>>
>>
>> -------------------------------------------
>> Richard Williams, Notre Dame Dept of Sociology
>> OFFICE: (574)631-6668, (574)631-6463
>> HOME:   (574)289-5227
>> EMAIL:  Richard.A.Williams.5@ND.Edu
>> WWW:    http://www.nd.edu/~rwilliam
>>
>> *
>> *   For searches and help try:
>> *   http://www.stata.com/support/faqs/res/findit.html
>> *   http://www.stata.com/support/statalist/faq
>> *   http://www.ats.ucla.edu/stat/stata/
>>
>
> *
> *   For searches and help try:
> *   http://www.stata.com/support/faqs/res/findit.html
> *   http://www.stata.com/support/statalist/faq
> *   http://www.ats.ucla.edu/stat/stata/



*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2021 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index