Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.


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

Re: st: theory reg vs. qreg


From   Kasal Roman <[email protected]>
To   [email protected]
Subject   Re: st: theory reg vs. qreg
Date   Tue, 30 Apr 2013 10:25:09 +0200

"qreg" is appropriate even for symmetrical distribution (but a real
data are not always perfect...almost never :) ) where it shows it's
more robust.

thank you

r

On Tue, Apr 30, 2013 at 10:18 AM, Yuval Arbel <[email protected]> wrote:
> -qreg- is the appropriate way for non-symmetrical distribution!!!
>
> Take a look at Greene's research I mentioned before:
>
> he ran the same model for different quantiles: the first, second, etc.
>
> On Tue, Apr 30, 2013 at 12:32 AM, Kasal Roman <[email protected]> wrote:
>> Yes, quantiles are not symmetrical estimates = very inconvenient when
>> there is a need to calculate complex survey data hypotheses.
>>
>> I thought it wouldn't be possible to treat  it (build and combine
>> equations) as with the "reg" command.
>>
>> Just tried if there would be any way for non-symmetrical estimates
>> ....probably there is not.
>>
>> thx
>>
>> On Tue, Apr 30, 2013 at 8:46 AM, Yuval Arbel <[email protected]> wrote:
>>> Roman,
>>>
>>> The feature you are referring to is the fact that the regression line
>>> passes via the sample mean.
>>>
>>> This is the reason why the projected Y for mean(X) is mean(Y).
>>>
>>> This outcome emanates from the derivation of the OLS formula, where we
>>> minimize the RSS (Residual Sum of Squares).
>>>
>>> As for qreg: it refers to a procedure where we minimize the absolute
>>> sum of deviation from  the regression line. It is also called median
>>> regression.
>>>
>>> The answer to your question is therefore depends on whether the
>>> distribution of the error term is symmetrical (in which case the
>>> median equals the mean) or skewed to the right or left.
>>>
>>> You should also take a look at the following textbook:
>>>
>>> William H. Greene: Econometric Analysis, Seventh Edition, p. 248-249.
>>> Greene describes his own research, which compares between estimates
>>> obtained from the median and OLS regressions
>>>
>>> On Mon, Apr 29, 2013 at 10:23 PM, Kasal Roman <[email protected]> wrote:
>>>> be a little constructive and don't talk about some mud, its obvious that:
>>>>
>>>> <<
>>>>>>sysuse auto
>>>>>>reg price weight length
>>>> you get the equation: Y1=a1*x1+a2*x2+b1
>>>>>>reg price
>>>> you get Y2=b2=mean(price)
>>>>
>>>> then (you can simply prove):
>>>>
>>>> Y1 = mean(weight)*a1+mean(length)*a2+b1 = b2 = Y2 = mean(price) = 6165.2
>>>>
>>>>
>>>> theory shows the same
>>>>
>>>>
>>>> On Mon, Apr 29, 2013 at 2:19 PM, William Buchanan
>>>> <[email protected]> wrote:
>>>>> If the editor of the Stata journal and a colleague of John Tukey are questioning your claims, I would venture to say:
>>>>>
>>>>> 1. What you claim is clear is about as transparent as mud.
>>>>> 2. Your notational conventions convey information in a way that is known only to you.
>>>>> 3. You want help but are unwilling to provide anyone with enough information to help you.
>>>>>
>>>>> The members on the Statalist come from a variety of disciplinary backgrounds and different countries, so what may seem clear as day to you can be clear as mud to others.  If numerous people respond to you with the same general line of thought (e.g., the way you are explaining things is unclear), then it is extremely likely that the problem is the way you are describing your problem.  You could help yourself quite a bit if you provided more explicit and detailed information like listserv members continue to request.  In the meantime, you may benefit from reading the Statalist FAQ and follow some of the guidance regarding how to develop questions that are likely to yield responses.
>>>>>
>>>>> Billy
>>>>>
>>>>> Sent from my iPhone
>>>>>
>>>>> On Apr 29, 2013, at 5:08, Kasal Roman <[email protected]> wrote:
>>>>>
>>>>>> Later I will put it here more detailed but its quite clear that on the
>>>>>> same dataset with a simple regression we get:
>>>>>>
>>>>>> Y1=Y2 when replaced "a2" for "mean(x)" in the equation
>>>>>>
>>>>>> thx
>>>>>>
>>>>>> r
>>>>>>
>>>>>>
>>>>>> On Mon, Apr 29, 2013 at 1:54 PM, David Hoaglin <[email protected]> wrote:
>>>>>>> Roman,
>>>>>>>
>>>>>>> What is the fact?  How is Y1 related to Y2?  Apparently, you have
>>>>>>> tried something with a -reg- command.  Please show us the command and
>>>>>>> the output.
>>>>>>>
>>>>>>> David Hoaglin
>>>>>>>
>>>>>>> On Mon, Apr 29, 2013 at 7:49 AM, Kasal Roman <[email protected]> wrote:
>>>>>>>> but this is a fact, just try it with the simple "reg" command:
>>>>>>>>
>>>>>>>> A) Y1 = b1
>>>>>>>> B) Y2 = a2*X + b2
>>>>>>>>
>>>>>>>> on the same data
>>>>>>>> then MUST be:
>>>>>>>>
>>>>>>>> Y2 = mean(x)*X + b2 = b1 = Y1
>>>>>>>>
>>>>>>>> and I'd like to know if there is something like this possible with the
>>>>>>>> "qreg" command.
>>>>>>>>
>>>>>>>> thx
>>>>>>>>
>>>>>>>> roman
>>>>>>> *
>>>>>>> *   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/
>>>>
>>>> *
>>>> *   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/
>>>
>>>
>>>
>>> --
>>> Dr. Yuval Arbel
>>> School of Business
>>> Carmel Academic Center
>>> 4 Shaar Palmer Street,
>>> Haifa 33031, Israel
>>> e-mail1: [email protected]
>>> e-mail2: [email protected]
>>>
>>> *
>>> *   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/
>
>
>
> --
> Dr. Yuval Arbel
> School of Business
> Carmel Academic Center
> 4 Shaar Palmer Street,
> Haifa 33031, Israel
> e-mail1: [email protected]
> e-mail2: [email protected]
>
> *
> *   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/


© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index