Amend life satisfaction scores to be in range 0-10 to align with ONS use of variable

[andrewbaxter439]

These scores are currently 1-7 which is how variable is measured in UKHLS. ONS life satisfaction variables measure on 0-10 score and wellbeing-adjusted life-years are calculated on this scale. Corrections currently in place to convert the 1-7 score when calculating wellbys but ideally 0-10 score should be used consistently?

[andrewbaxter439]

Hi @dkopasker - could you pop in stata code for the 1-7 -> 0-10 conversion for life satisfaction? I think it would be something like: (lf7 - 1) * 10/6), giving the mapping:

lf7	lf10
1	0.00
2	1.67
3	3.33
4	5.00
5	6.67
6	8.33
7	10.00

[dkopasker]

Below is the code used in the Comparative Panel File to rescale life satisfaction on a ten-point scale. lfsato comes from BHPS, sclfsato is the equivalent variable in Understanding Society. Note that satlife10 will not adhere to the new variable naming structure. You will also want to handle the missing values differently, i.e. by setting to ".". However, the alternative approach you suggested seems much simpler.

recode lfsato (1=0)(2=1.67)(3=3.33)(4=5)(5=6.67)(6=8.33)(7=10) ///
(-9 -2 -1=-1)(-8 -7=-3), gen(temp_a10)
recode sclfsato (1=0)(2=1.67)(3=3.33)(4=5)(5=6.67)(6=8.33)(7=10) ///
(-10 -9 -2 -1=-1)(-8 -7=-3), gen(temp_b10)
gen satlife10=temp_a10
replace satlife10=temp_b10 if wave>=19
drop temp_*

[andrewbaxter439]

Closed by #307