tab2 harvester demonstration
will give you the mix of the two
will give you the mix of the two
-
Login or Register
- Log in with
- You are not logged in. You can browse but not post. Login or Register by clicking 'Login or Register' at the top-right of this page. For more information on Statalist, see the FAQ.
-
-
George Ford Thank you for your reply.
Yes. All demonstrations are on harvester use (i.e., the benefits of using harvester). However, it is possible that all farmers who attend the demonstrations may not actually adopt the harvester.
Comment
-
George Ford Thank you for your reply
Yes. All demonstrations show and motivate farmers on the use of the harvesters. So there is no demonstration without the harvestor. However, it is possible that not all farmers who attend the demonstrations will adopt the harvester.Comment
-
Yes all demonstrations show and motivate farmers on the use of the harvester. SO not demonstrations without a harvester. However, it is possible that all the all the farmers who attend the demonstration do not adopt the harvester.Comment
-
then you have 2 treatments, even though I would not expect a demo without a harvester used after has any effect.Comment
-
Does this mean that I should construct the treatment variable based on demonstration and adoption of harvester?
So treatment = 1, if in treated group and adopted harvester?Comment
-
treatment0 = demonstration
treatment1 = harvester use (since all harvester use has a demo, this is fine)
treatment0*post gives the effect of the demo
treatment1*post gives you the effect of the harvester
should be no effect of demo without harvester use.
To get the effect of the demo, you need harvester use without a demo, which you don't have. You are limited to testing the effect of the harvester.Comment
-
-
the demo treatment might indicate selection bias (if it's non-zero, then the demo may have been given to better or worse than average farmers).Comment
-
So I would need additional tests that control for selection bias?
Such as treatment effect models?Comment
-
the demo*post variables should tell you if there's an issue.
if they are significant, then you might think about a selection model of some sort.Comment
-
I get it.
Thank you.Comment
-
Dear Prof. George Ford,Originally posted by George Ford View Post
I would like to ask a bit about my specification. Specifically, I want to examine the effect of a specific law which was in effect in 2008. That means estimates are intention to treat. My simple econometric model is as follows: y = alpha0 + alpha1*law +alpha2*control + u, where y is the outcome of interest and law is an indicator taking a value of 1 if post-law periods. Here are my questions:
1) I have three waves of repeated cross-sectional data in 2006, 2010 and 2012. Can I call the equation above a DID estimation?
2) Since I have only one pre-law data point, how should I test for pre-trend assumption?
Thank you.
Comment
-
If you have at least one pre-law period, you can use DID, as long as some units are not treated in any period.
You can't see the pre-trend with only one period. You might find something similar to the outcome of interest for which data is available and see if it followed a similar trend.Comment
-
Dear Prof. George,Originally posted by George Ford View Post
Thank you for your response and advice.
In my setup, individuals surveyed in 2006 were not treated, but I am not sure if they were reinterviewed in 2010 and 2012 given the nature of repeated cross-sectional data. In this case, can I use the equation in #28 to estimate the effect of the law? it seems that I have a before and after setup.If you have at least one pre-law period, you can use DID, as long as some units are not treated in any period.
Could you please elaborate this point a bit? with an example would be great.You might find something similar to the outcome of interest for which data is available and see if it followed a similar trend.
Thank you.Comment
Comment