Create a Pseudo Panel Data and Analysis

[GBEHE Bienvenu]

Hello eveyone

I would like to analyze the determinants of the dynamics of women's employment (15-49 old) in the Urban and Rural Sector in Ivory Coast : Pseudo Panel Approach

I want to create a pseudo panel dataset, I have 3 surveys datasets with year of interview 1999 and 2005 The three surveys have a some variables.

use "data1990.dta, clear
egen scohort = concat(urban age_group), decode punct("_") // create a stable cohort by age_group and residence zone

foreach var of varlist work child educ tv phone married ethni relig {
bys year scohort: egen avg_`var' = mean(`var') 
}  // create a mean of each variable

save base1.dta, replace

use "data2005.dta, clear
egen scohort = concat(urban age_group), decode punct("_") // create a stable cohort by age_group and residence zone

foreach var of varlist work child educ tv phone married ethni relig {
bys year scohort: egen avg_`var' = mean(`var') 
}  // create a mean of each variable

save base2.dta, replace
append using base1.dta
save base3.dta, replace

I understood that in the pseudo panels, we follow the generation and not the individuals.

So i have created la variable generation ou (ID) like that : people aged 15 to 19 in 1999 will be between 21 and 25 in 2005 and 30 to 34 in 2012. this gives us generation 1 follow-up over all years. with their observation in average

gen gen_co= 1 if (age>=15 & age<=19 & year==1999) | (age>=21 & age<=25 & year==2005) | (age>=30 & age<=34 & year==2015)

replace gen_co = 2 if (age>=20 & age<=24 & year==1999) | (age>=26 & age<=30 & year==2005) | (age>=33 & age<=37 & year==2015)

replace gen_co = 3 if (age>=25 & age<=29 & year==1999) | (age>=31 & age<=35 & year==2005) | (age>=38 & age<=42 & year==2015)

however when I did the analysis

xtset gen_ID year 
repeated time values within panel

Then I did

xtset gen_ID
xtreg avg_work avg_child avg_educ avg_tv avg_phone avg_married avg_ethni avg_religion age agesq, fe
estimates store FE1
xtreg avg_work avg_child avg_educ avg_tv avg_phone avg_married avg_ethni avg_religion age agesq,re 
estimates store FE2
esttab FE1 FE2

I would like to know if my understanding and steps are good

[Williams Yaji]

Hello Gbehe, I would like to know how you proceeded with the analysis. I am currently undertaking a similar study but I'm stuck.