I am trying to enumerate the effective sample size of sub-samples of a large
scale survey using only the relative weight of each respondent.
I have several files with around 1000 observations in each and their
relative weights. Some respondents have their own unique weight but others
share the same weight with a handful of other respondents. I am treating
each different weight value as a stratum and then, to calculate the
effective sample size, using the formula:
sum across h of (nh)*(wh) - all squared
divided by the sum across h of (nh)*(wh)squared
where (nh) is the sample size in stratum h and (wh) is the weight applied to
sample elements of stratum h.
The results I am getting are much lower than I would expect (around 20!).
The weights typically vary from 0.3 to around 4, but most are not too far
from 1.
Can I ask if you are able to verify the formula I am using or suggest an
alternative approach?
Many thanks in advance,
Karsten
### This e-mail and all attachments it may contain is confidential and
intended solely for the use of the individual to whom it is addressed. Any
views or opinions presented are solely those of the author and do not
necessarily represent those of Ipsos UK and its associated companies. If you
are not the intended recipient, be advised that you have received this
e-mail in error and that any use, dissemination, printing, forwarding or
copying of this e-mail is strictly prohibited. Please contact the sender if
you have received this e-mail in error. ###
|