Creating reliable fertility history variables - UKHLS waves 1-8
Dear UKHLS team
I have had a look at the existing FAQs but haven't been able to find a response, but please let me know if I'm incorrect.
I am using UKHLS waves 1-8 (which I have merged) and want to do the following:
(1) create a variable that flags all individuals that have had their FIRST child in wave 4
(2) create a variable that flags all individuals that have remained childless for all waves, 1-8
(1) and (2) make up the two groups I am comparing/interested in.
--> (i) *_nnatch : number of natural children in the household
--> (ii) *_lprnt : whether have ever had any children (new respondents only AND if answer to *_nnatch is zero)
--> (iii) *_nnewborn : number of children had since last interview (repeat respondents)
--> (iv) xw_anychild_dv : whether have ever had a child
(where the asterisk denotes a variable common across all waves, with wave prefix a, b, c etc.; xw denotes it's an xwave variable).
For (1): am I correct in thinking I need to create, first, a variable that can confirm that an individual has remained childless from wave 1 to wave 3 (using nnatch, lprnt and nnewborn) and subsequently had a child at wave 4 (using nnewborn).
For (2): would the xw_anychild_dv variable be sufficient?
As regards the latter point, any information on how this derived variable is calculated would be very helpful (aside from the fact it is calculated from ch1bm and ch1by). The reason I ask is that when I check the consistency of this variable I don't understand the results. So for example, if I ask STATA "tab d_nnatch if anychild_dv==2", the results tell me there are 72 individuals that had more than 0 children in their household at wave 4, despite the xwave variable suggesting they have never had children (anychild_dv==2).
Thank you very much in advance and kind regards