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: ATT greater than one


From   Isobel Williams <[email protected]>
To   "[email protected]" <[email protected]>
Subject   Re: st: ATT greater than one
Date   Sun, 16 Mar 2014 14:47:43 +0000

I am measuring the likelihood that a child would enroll in school as the result of the welfare programme using a logit model. After matching, the ATT is always less than zero, so I interpreted the change as percentage points. 

However, when I then use the same technique to measure the ATT on number of hours the child works between the treated and the untreated, I get extremely high numbers for the ATT, some in the thousands. I was wondering how to interpret these results. 

Thanks,
Isobel


> On Mar 16, 2014, at 2:39 PM, "Ariel Linden" <[email protected]> wrote:
> 
> You haven't provided us with sufficient information to respond to your
> question. Why would you think that ATT should be interpreted as percentage
> points??? It all depends on what your variables are and what it is you're
> estimating. If you were estimating medical costs as an outcome, then these
> results may be very reasonable (since medical costs typically follow a
> Poisson-like distribution). 
> 
> 
> 
> Date: Sat, 15 Mar 2014 19:34:24 +0000
> From: Isobel Williams <[email protected]>
> Subject: st: ATT greater than one
> 
> Dear Statalist,
> 
> I'm using psmatch2 (user-written software) to find the ATT of a welfare
> programme. In some of my results, the ATT is coming up in extremely high
> numbers. The standard errors are also really high.
> 
> Does this mean that I've done something wrong? How would I interpret this?
> 
> I know that ATT should be interpreted as percentage points, i.e. an ATT of
> 0.67 would mean an increase of 6.7 percentage points, but what about an ATT
> of 1482.47, with a standard error of 1497.69?
> 
> Many thanks,
> Isobel Williams    
> 
> *
> *   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