Understanding Society User Support

Hi Xavier,
the differences between the two variables are indeed minute. The difference stems from a slight difference in the way that the variables qfachi and qfedhi (in the stand-alone BHPS) deal with qualifications reported by proxy. If the proxy is present again in indresp in the following year (either as a proxy or full respondent) their previous wave value on qfachi and qfedhi will be considered in computing the current wave qfedhi and qfachi. However, if the proxy misses one wave and returns in a future wave, their previous wave value is not considered. bw_hiqualb_dv is a simple recode of qfachi and qfedhi so will have this problem. ba_hiqual_dv uses qfachi and qfedhi in hBHPS wave 1 feeds it forward to wave 2, where bb_hiqual_dv is created from bb_qfedhi and bb_qfachi and ba_hiqual_dv, whichever is the lower value (i.e., higher degree). Then, in wave 3, both ba_hiqual_dv and bb_hiqual_dv are feed forward and evaluated against bc_qfedhi and bc_qfachi to create bc_hiqual_dv. At this stage the first differences between bc_hiqualb_dv and bc_hiqual_dv occur.

Hope this explains,
Gundi