I am trying to build a nested loop setup, which can be used to generate mode of responses. The loops are on treatments and individual IDs. Furthermore, I have to put in some conditions to filter out certain set of data. The conditions are on gender, individual ID and treatments. Within each treatment, I have to exclude the individual for whom I am generating the mode, then I have to filter for gender, either male or female. Then I have to randomly select 5 person from those who remains. Then calculate the modal response of them for the question. I am currently using the following code but it is generating missing values for multiple individuals across multiple treatments. The number is increasing as I am moving towards later iterations.
The code I am using is as follows -
set seed 12345
gen random = runiform()
forvalues t = 1(1)5 {
forvalues i = 0(1)4 {
foreach j in u m c d{
* Reload the original dataset for each combination
use "datafile1.dta", clear
* Filter for current treatment
keep if treatment == `t'
* Generate individual ID
gen t`i'`j'indv_id = _n
* Get list of IDs
levelsof t`i'`j'indv_id, local(idp)
foreach k of local idp {
* Reload data for each ID to avoid cumulative filtering
use "${path}/Temp/datafile1.dta", clear
keep if treatment == `t'
gen t`i'`j'indv_id = _n
* Apply filters
keep if t`i'`j'indv_id != `k'
keep if p3q2_gender == 0
sort random
keep in 1/5
* Calculate frequency and mode (corrected syntax)
bysort p1q3_t`i'`j'_s : gen freq_p1q3_t`i'`j'`k' = _N if !missing(p1q3_t`i'`j'_s)
egen mode_p1q3_t`i'`j'`k' = mode(p1q3_t`i'`j'_s), maxmode
summarize mode_p1q3_t`i'`j'`k'
scalar p1q3_t`i'`j'`k'_mode = r(mean)
* Save and append
drop freq_p1q3_t`i'`j'`k'
tempfile temp_t`i'`j'`k'
save `temp_t`i'`j'`k''
use "${path}/Temp/clean1.dta", clear
append using `temp_t`i'`j'`k''
save "${path}/Temp/clean1.dta", replace
}
}
}
}
The used datasets / dta files are already created beforehand. Also after I get the mode, I want to save it as a scalar and associate it with the individual for whom it was generated.
Every treatment has more than enough observations (around 160 and almost equal number of male and female). I am not getting why is it generating missing values and then no modal value.
If anyone can help me or find the error in the logic I am using, it would be a great help.
Thank You.
The code I am using is as follows -
set seed 12345
gen random = runiform()
forvalues t = 1(1)5 {
forvalues i = 0(1)4 {
foreach j in u m c d{
* Reload the original dataset for each combination
use "datafile1.dta", clear
* Filter for current treatment
keep if treatment == `t'
* Generate individual ID
gen t`i'`j'indv_id = _n
* Get list of IDs
levelsof t`i'`j'indv_id, local(idp)
foreach k of local idp {
* Reload data for each ID to avoid cumulative filtering
use "${path}/Temp/datafile1.dta", clear
keep if treatment == `t'
gen t`i'`j'indv_id = _n
* Apply filters
keep if t`i'`j'indv_id != `k'
keep if p3q2_gender == 0
sort random
keep in 1/5
* Calculate frequency and mode (corrected syntax)
bysort p1q3_t`i'`j'_s : gen freq_p1q3_t`i'`j'`k' = _N if !missing(p1q3_t`i'`j'_s)
egen mode_p1q3_t`i'`j'`k' = mode(p1q3_t`i'`j'_s), maxmode
summarize mode_p1q3_t`i'`j'`k'
scalar p1q3_t`i'`j'`k'_mode = r(mean)
* Save and append
drop freq_p1q3_t`i'`j'`k'
tempfile temp_t`i'`j'`k'
save `temp_t`i'`j'`k''
use "${path}/Temp/clean1.dta", clear
append using `temp_t`i'`j'`k''
save "${path}/Temp/clean1.dta", replace
}
}
}
}
The used datasets / dta files are already created beforehand. Also after I get the mode, I want to save it as a scalar and associate it with the individual for whom it was generated.
Every treatment has more than enough observations (around 160 and almost equal number of male and female). I am not getting why is it generating missing values and then no modal value.
If anyone can help me or find the error in the logic I am using, it would be a great help.
Thank You.
