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: Specification text for continuous Interaction term model over dummy variable model.
From
Nick Cox <[email protected]>
To
[email protected]
Subject
Re: st: Specification text for continuous Interaction term model over dummy variable model.
Date
Thu, 19 Jan 2012 11:48:17 +0000
You could add both terms to a wider model.
Nick
On Thu, Jan 19, 2012 at 11:23 AM, natasha agarwal
<[email protected]> wrote:
> Thanks John,
>
> But is there a statistical test to formally test the same?
>
> Thanks,
> Natasha
>
> 2012/1/19 Johan Hellström <[email protected]>:
>> Natasha,
>>
>> I would say the first model, the continuous*continuous interaction model, as in model 2 you throw away a lot of potentially useful information in the data.
>>
>> Best,
>> Johan Hellstrom
>>
>>>Dear All,
>>>
>>>I was trying to estimate a model first with continuous interaction
>>>term and then as a dummy model, and I wanted to know which one would
>>>be my preferred model. Is there any test to test the same.
>>>
>>>Model 1:
>>>
>>>xtreg x y z a*b, fe vce(robust) [a*b=continuous interaction term]
>>>
>>>Model 2:
>>>
>>>xtreg x y z a*c, fe vce(robust) [c=1 if b>0.50]
>>>
>>>I want to test which is a better model?
>>>
*
* 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/