Hi Sweety,

Since there are two time periods (according to Card,1992 setup), it should

be differenced over time (the difference between the two).

Hope this helps

 Thank you sir for your valuable insights.

So should I subtract  the values of  pre- treatment year from post treatment year or  the values of post treatment year from pre-treatment year or in any direction can I estimate the difference?

Thanks Sweety.

It should be the difference/change in values(Delta).  If you are familiar with STATA, then you can use their code.  

Hope this helps

So all the independent variables and error term should be changed accordingly right?

Hi Sweety,

What  are the 2 groups in your model for which you are using the DD method?

if you let me know a bit more about your groups, then I can advise on the model, constructing variables...

Sir, My treatment group is National Food Security implemented states in India and control group is Non-NFSA implemented group in India. My outcome variable is average quantity of rice consumed.

Hi Sweety,

To do this, you can follow the following strategy.

1. Assume you have:

Group 1: National Food Security implemented states in India (call it group A)

Time period 1(T1)

Group 2. Non-NFSA implemented (call it group B)

Time period 2(T2)

2. Generate a control group dummy call it 'Sweety' which is equal to one if a company/person is in group A and zero otherwise, generate a time dummy call it 'time', which is equal to one if t=2 and zero otherwise

3. Create a dependent variable call it Yit (i.e the outcome variable, company performance for example) use the variable as is without difference (‘i’ refers to a person or company based on your data; t is the time ‘year or month’ according to your data), and then regress

Yit=α+β1(sweety)+β2(time)+ρ(sweety⋅timet)+ϵit

And then explain your coefficients/ results

Hope this helps,                    

Thank You Sir. I have used the same equation. But as you told before Yit must be delta (change in) Yit right?

Hi Sweety,

No, according to this strategy you should use the dependent variable as is (without difference or delta). There is another strategy for which you have to use the difference in the dependent variable (delta), but it is a bit complicated and time consuming.

Hope this helps and good luck with your model.

Thank You Sir. So I hope that I can take quantity in pre and post period as dependent variable and can compare the changes during both the years.

Good luck with your estimations. Happy to answer any further questions.

Thank you Sir