Understanding Society User Support

Hi Piotr

Many thanks for this. I am still puzzled by some of the participants who have ‘not applicable’ for i_macurr.

I have restricted the sample to foreign born participants only (using the variable bornuk_dv from the xwavedat dataset – I deleted all UK born participants and participants with missing data for this variable).
Of the remaining participants, I then removed participants whose mother was born in the UK (using the variable macob from the xwavedat dataset – I deleted all participants with UK born mothers and participants with missing data for this variable)

Of the then remaining participants, I then have removed participants who entered the study at wave 6 or earlier and were in the IEMB sample ( using the fwenum_dv variable from the xwavedata dataset and the i_iemb from the i_hhresp dataset)

I still find that 591 (17.72%) of the remaining sample have the value ‘not applicable’ for i_macurr.

Have I used the correct versions of the variables to restrict the sample and/or is there another reason participants may have ‘inapplicable’ for this variable?

Many thanks
Rachel

Hi Rachel,

What statistical package do you use?

Thanks,
Piotr
UKHLS User Support

Hi Piotr

I use STATA.

Thanks
Rachel

Hi Piotr

Sure, here it is :

*CREATING FILE WITH 'I_fborn' and 'i_iemb' VARIABLES TO MERGE INTO MAIN FILE**
use "UKDA-6614-stata\stata\stata13_se\ukhls\i_hhresp.dta"

keep i_hidp i_fborn i_iemb
sort i_hidp
save fborn, replace

CREATING FILE WITH STABLE CHARACTERSITICS, INCLUDING WHERE MOTHER/FATHER WAS BORN***
use "UKDA-6614-stata\stata\stata13_se\ukhls\xwavedat.dta"

keep pidp sex_dv bornuk_dv plbornc yr2uk4 generation macob pacob anychild_dv fwenum_dv

sort pidp

save stablecharacteristics, replace

*MAIN FILE (USING WAVE 9 INDIVIDUAL RESPONDENTS)**
use "UKDA-6614-stata\stata\stata13_se\ukhls\i_indresp.dta"

sort pidp

MERGE IN STABLE CHARACTERISTICS FILE
merge pidp using stablecharacteristics
tab _merge
drop if _merge==2
drop _merge

tab bornuk_dv, m

***deleting cases with missing data for bornuk_dv (588, 1.63% of sample)
drop if bornuk_dv==-9

***DELETING CASES WHERE PARTICIPANTS WERE BORN IN THE UK
drop if bornuk_dv==1

deleting cases where mother was born in the uk or missing data for mother's country of birth
drop if macob<5

MERGE IN FILE WITH FOREIGN BORN/ETHNIC MINORITY BOOST SAMPLE VARIABLES*
sort i_hidp
merge i_hidp using fborn
tab _merge

this deletes some cases here which do not hhresp file
keep if _merge==3

deleting cases which entered at wave 6 or earlier AND were part of IEMB sample *
drop if fwenum_dv<7 & i_iemb==1

*distribution of i_macurr*
tab i_macurr

Many thanks,
Rachel

Thank you. We'll look into it.

Hi Rachel,

A quick update: I can confirm that there might be an issue with the macurr routing, we are investigating possible reasons. I will let you know when I learn more.

Best wishes,
Piotr

Many thanks Piotr

Hi Piotr,

Is it possible to provide a timescale for this? I have some analysis due by the end of this week which I was hoping to use this variable in.

Many thanks
Rachel

Hi Rachel,

We have conducted an internal investigation to identify any patterns behind this error but have not found a consistent explanation for all the mismatches. Here's what we discovered:
a. 200 individuals were not asked i_MACURR, even though they should have been. All of these cases were part of the Quarter 2 sample.
b. 414 individuals were asked i_MACURR, even though they shouldn't have been. Of these, 261 were from the Quarter 3 sample, with the remainder spread across the other seven quarters.

Given this, we recommend filtering out the cases in group (b) using the universe condition. (you can use the syntax you shared with us earlier)

We will keep you informed if we learn any further details about the cause of this filtering/routing issue.

I hope this information is helpful.

Best wishes,
Roberto Cavazos
Understanding Society User Support Team

Hi Roberto

Many thanks for looking into this - that's really useful information and I will follow you strategy you propose.

Thanks
Rachel