Support #216

Why so many deaths in 2000 and 2004?

Added by Stephen Clark about 10 years ago. Updated about 10 years ago.

Start date:
% Done:




I am just starting to get to grips with the BHPS and I have a question as to why there were so many deaths recorded in 2000 and 2004. I am looking at the wIVFIO=99 variable/value in the SPSS file xwaveid.sav. The number of deaths in 2000 is 232 and in 2004 it is 293 – these values are just less than twice the usual “trend” values for neighbouring years.

On this forum I can only work out how to see new/in progress threads (14 of them) but there is a counter that suggests there are actually 190 threads. If this has been asked in one of the threads I can't get to see, I apologise. I have seen the open post by John MacInnes and his link to a discussiong paper by Lui and Weale which I shall read.


Updated by Redmine Admin about 10 years ago

  • Status changed from New to In Progress
  • Target version set to BHPS
  • % Done changed from 0 to 50

The survey office put more effort into verifying deaths of sample members in certain years if the sample members or their households had not been contactable for one or more years. This was done at the survey office by searching obituaries and stressing the importance of gathering this information to interviewers before the fieldwork. If you look back at the household interview outcomes of deceased sample member a wave or two you will see a varying proportion originates from households that were unproductive in the previous year. This effect is particular clear for Wave 14. The new Scotland and Wales samples were first selected at Wave 9 and the mortality rate in these two samples seems to be higher than average for some reason in the following year.


Updated by Redmine Admin about 10 years ago

  • Status changed from In Progress to Closed
  • % Done changed from 50 to 100

Also available in: Atom PDF