Understanding Society User Support

Hello,

I am using Wave 13 data for my analysis and have been looking to create household or HRP specific variables as part of this (using the xwave file to link individuals and their households, whilst allowing for additional waves to be added for later analysis). However, during this process I have noticed some inconsistencies where some ~4000 households seem to have children listed as their HRP in the household level data rather than an adult. I noticed this when creating a flag variable to more easily identify HRPs which provided value of 1 when the pidp of an individual matched the identifier for the HRP of the household (using the hrpid variable). Other household variables (notably hidp) are otherwise consistent between individuals within each household.

I also wished to check that my approach to correcting this potential error was correct for the dataset. I have derived a new HRP identifier which first identifies who in each household owns/rents the property the household is in (using both the rentp and hsowr1 variables and looking to see if these variables matched the pno of each individual in the household) and then, if multiple people are responsible, assumed the oldest of them was responsible (which seemed to follow the logic listed on the UKHLS website for the hrpid variable). In cases where multiple people are responsible and the same age, I assumed the first person listed (using pno values) was appropriate. In addition, for households where no individuals are listed as responsible (or where this person did not complete the individual questionnaire), I assumed that the first adult who completed the individual questionnaire was the HRP for their household.

Any feedback on this potential issue would be greatly appreciated and many thanks for your help!

Dan Pinhay