Congratulations to John Barker both for bringing this list to life and for
initiating a discussion of population projections.
But, as the discussion already indicates, anyone who treads into this area
is liable to be subject to a flood of simplistic opinion, abuse and
misrepresentation. Yet the statistical background to matters of
population is usually more robust and reliable than that of any other
subject area.
The term 'usually more reliable' is used because the 2001 Census results
have thrown a lot of uncertainty into the situation. Paul Spicker is quite
right to emphasise the administrative importance of the Census. It is
administratively very inconvenient to deal with real world complexities than
cannot be summarised by single numbers. That was the justification for the
One Number procedures that delayed publication by 18 months. But it not
quite right to say, as Paul Spicker says, that:
>The problem with the net migration estimates, then, is not
>that they have
>been too low, but that they have been too high.
The problem may rather have been with the Census. The migration estimates
were revised to fit the results of the One Number Census. The ONS, having
committed itself to a One Number Result, then found that it faced a major
problem. The response rate from the Census was lower than ever before.
But the response rate from the Census Coverage Survey was even lower!
So how to get One Number? Only by statistical imagining! Statistical
theory postulates that there is a true value for the total population. The
ONS assumed that the Census and the CCS provided independent estimates
related to that true value. Belief in this theory and the assumption of
independence supported the 58,789,194 estimate. To believe the 58,789,194
result, you have to believe that true values exist. To believe the
58,789,194 result, you also have to believe that the Census and the Census
Coverage Survey did not suffer from similar biases.
It is easier to believe that estimates are the product of following specific
data collection and production procedures.
The One Number Census seems to be a manifestation of the English sin of
hypocrisy in the area of statistics. Let us pretend that there is no
migration problem, that there are no illegal immigrants, that everyone wants
to be counted!
I can imagine that one or two readers of this list are already foaming at
the mouth at the use of the phrase 'illegal immigrant' that is not seen as
politically correct in some circles. But it is worth noting that our North
American 'ally' does have a rather larger illegal immigrant problem, does
acknowledge the existince of the problem, and has even devised methods of
measuring the number of illegal immigrants without requiring revelation of
they are living.
Ray Thomas, Social Sciences, Open University
Tel: 01908 679081 Fax 01908 550401
Email: [log in to unmask]
35 Passmore, Milton Keynes MK6 3DY
******************************************************
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]
*******************************************************
|