st: Question about comparing sqreg models

[Milena Nikolova <mnikolova01@gmail.com>]

Hi Billy and James: 

Thank you both for your responses. I am trying to compare the coefficient estimates (mainly significance) between two models - one with, say, 10 controls (Model 1), and one with 13 controls (Model 2). In other words, I'd like to compare what happens to the coefficients estimates of Model 1 evaluated at, say, the median, when I add additional controls. Does this make more sense?

@James: So, are you suggesting that the two models can be compared just like OLS models? My main concern was that some models e.g., logits/probits/ologits cannot be compared the same way that OLS models can (i.e., using the same techniques and tests) as Richard Williams, among others, has so eloquently explained in this Statalist and elsewhere. I was wondering whether there is anything about the error variance or anything else about quantile regressions that make comparisons across models incorrect. 

Many thanks in advance!!!
Milena




On Aug 21, 2013, at 4:27 PM, William Buchanan wrote:

> There may be some confusion regarding the way you describe the models.  When you refer to the first model as the "truncated model" it isn't clear if you're talking about a truncated distribution or if you are just describing a simplified version of your final model.  What results are you attempting to compare between the models?  
> 
> HTH,
> Billy
> P: 401-499-9719
> 
> 
> 
> On Aug 20, 2013, at 6:42 PM, Milena Nikolova <mnikolova01@gmail.com> wrote:
> 
>> Hello!
>> 
>> I am using Stata 11.2 for Mac and I have two regression equations (the dependent variable in each is the same and is an ordinal variable measured on a scale from 0 to 10) estimated using -sqreg- for the 10th, 25th, 50th, 75th, and 90th percentiles. Model 1 is the truncated model and Model 2 is the full model (with additional controls). My questions are: i) what is the methodologically appropriate way of comparing the results from Model 1 with those from Model 2? ii) What command could I use in Stata to get at this comparison?
>> 
>> Many thanks in advance!
>> 
>> Milena
>> *
>> *   For searches and help try:
>> *   http://www.stata.com/help.cgi?search
>> *   http://www.stata.com/support/faqs/resources/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/faqs/resources/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/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/