This discussion has come back to what I consider one of the prime
functions of a good statistician, and one of the reasons why other people
find us boring and iritating. A *good* statistician has an obsessive
interest in knowing what the figures truly mean. (A really good
statistician can engage the client's interest, hopefully to reduce the
irritation.) What seems to have gone wrong at the start of the
presentation is an assumption that a "one figure" census is like a game of
statues - or the start of "In town tonight": the figure should be a simple
integer based on the number of bodies at one instant.
But even that definition can not work. How close to death - or birth -
should be included or excluded? Are emigrants included on the assumption
that they will return within a certain time? The discussion has mentioned
missing or overlooked people, but not those double-counted.
One concept I try to introduce to students very early on is that some
measurements are inherently fuzzy, so variation represents uncertainty
that is not "error" in any sense that can be "improved on". For planning
purposes, it seems to me more useful to have given lower and upper
estimates of the population rather than place all the emphasis on a claim
for a single "best unbiased estimate".
R. Allan Reese Email: [log in to unmask]
******************************************************
Please note that if you press the 'Reply' button your
message will go only to the sender of this message.
If you want to reply to the whole list, use your mailer's
'Reply-to-All' button to send your message automatically
to [log in to unmask]
*******************************************************
|