SF36GHP in Wave 9 of the BHPS
Since the input variables for the summary scales of the SF36 are available in Wave 9 and 14 of the BHPS, but the derived summary scales only in Wave 9, I have used the calculations supplied in the documentation to create the derived summary scales for Wave 14 too, validating them by re-creating those for Wave 9 and checking they match with your own.
This works fine for all but one derived variable - the SF36GHP - which in my version from the calculations increases in steps of 5, but in your derived one, from 10 to 12, 25 to 27 etc. But if you back calculate from these values then a value of say 97 would have to have been calculated from a combined score of 24.4 - not possible when the scales of each of the input variables are in whole integers. Can someone check what is going on with this? As I am tempted to use my version as it makes more sense, but as all the other derived variables come out right, I wondered if I'd missed something important!