Let y(G) denote the mean outcome for group G. The decomposition of the
difference in differences
D = [y(isr_nat) - y(isr_imm)] - [y(ger_nat) - y(ger_imm)]
is then computed as
. jmpierce2 ger_nat ger_imm isr_nat isr_imm
Germany is the benchmark. If you want Israel be the benchmark, type
. jmpierce2 ger_nat ger_imm isr_nat isr_imm, benchmark(2)
The choice of the benchmark does not change the overall differential.
However, it changes how the differential is decomposed. See the
formulas in -help jmpierce2-.
ben
On Mon, Jul 20, 2009 at 10:59 AM, Michael
Buryak<[email protected]> wrote:
> Hello, Statalist
>
> Something not clear for me in the results after using the module JMPIERCE2.
>
> For example:
> I have the estimated wage for four groups : two in Israel and two in Germany.
> In each country one group is natives and second group is immigrants.
> So four groups:
> 1) ISR_NAT
> 2) ISR_IMM
> 3) GER_NAT
> 4) GER_IMM
>
> If I want to decompose difference in difference:
>
> D = [(D ISR_NAT - D ISR_IMM) - (D GER_NAT - D GER_IMM)]
>
> Is it my command right:
>
> reg wage hour age educ if isr_nat==1
> est store isr_nat
> reg wage hour age educ if isr_imm==1
> est store isr_imm
> reg wage hour age educ if ger_nat==1
> est store ger_nat
> reg wage hour age educ if ger_imm==1
> est store ger_imm
> jmpierce2 isr_nat isr_imm ger_nat ger_imm, benchmark(1)
>
> Which order of estimators in command jmpierce2 is right for my decomposition?
> If I change to the parameter benchmark(2), the total difference is same. Why?
>
> What is the role of parameters "benchmark" and "reference" in my order
> of estimators: isr_nat isr_imm ger_nat ger_imm?
>
> Thank you in advance for your help,
> Michael
>
> *
> * 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/
>
*
* 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/
