As someone who spends his life being part statistician and part journalist,
I found the shot at journalists (who presumably are not members of the list
and get no right of reply) a bit cheap. The lead article in the latest RSS
mag was speculating as to which of the two 'professions' had the lowest
reputation in society? I'm not sure which wins.

BUT so far as the substance goes ...... The scandal about the census is
surely the leap of faith that the three registrars are asking us to make -
and not a few unfortunate newspaper quotes.

As I see it, the 91 census "missed" (to quote the Treasury minister in the
2001 debate in the HoC) as many as 20% of young men in cities. The ONS
(OPCS as it was) added 000's to the numbers actually counted to get to the
total population. The 2001 census also 'missed' (my word) lots of young
men. This time, however, the ONS has decided that all the people are abroad
and have not added them back in. They also hint that many were abroad in
1991 but not in 1981. Cracking!

If there was any decent evidence to support the scale, timing and direction
of this net migration it would be easier to swallow. The scale of migration
is matched only by the scale of the holes in the data. The gross flows
overseas must be enormous - I can't remember the last time I was served by
a native English speaker in a London bar or restaurant.

Lets turn to the mid-year population estimates against which the census
totals can be compared. The strategic decision about migration trends,
coupled with apparently dodgy immigration figures (immigrants have been
counted into the inner cities, we are told, but not counted when they left
to go elsewhere) mean that some LAs have suffered large population cuts in
the census compared to the latest mid year estimate. The cut was 25% in the
case of Westminster. Just pause and think again - one LA has had its
ONS-provided measure of population cut by a quarter over night!!! Is this
the largest ever ONS revision of data? There are number of others where the
fall is in double digits. If the ONS can change the population of an area
by 25% it surely gives rise to a range of questions of credibility.

I would have liked more information about the CCS. (There are according to
the ONS dozens of research papers on the web if you go hunting for them but
that is not all that could be done.) It is not easy for generalists (such
as journalists and LA employees) to grasp the CCS and those in the ONS who
have been working on it for the last few years are - understandably perhaps
- so into it that they cannot explain it to outsiders. And they treat naive
questions as criticisms - I suppose what looks convincing in a formula
sometimes is less so in plain English.

As the Treasury minister explained in the debate "by using the findings of
the coverage survey in combination with the findings from the census, the
characteristics of these groups that were missed can be imputed". Maybe.
But I  must confess to having some difficulty explaining to generalists how
a census enumeration rate of say 70% and CCS rate of 50% means that you can
accurately impute the 30% you missed. Surely there is a very good chance
that those missing from the former are not well represented in the latter.
The ONS has told me that they do not have time to go through any worked
examples with me, so I am left wondering how exactly this works on a
deprived and dangerous inner city estate (of which there are many in
Westminster). The apparent correlation between low enumeration rates in LAs
and population cuts does not help to convince a sceptic.

There are many people in the world who believe in God and many who feel
just as strongly that there is no God. I'm in neither camp. The census
results seem to deliver a similar split - there is no proof one way or
another and the two 'sides' find it difficult to have a sensible
discussion. As it has taken 18 months to publish the results and it is a
world class census, I guess we ought to believe the numbers. Or should we?
Poor journalism might not help but this is a tricky one to sell.

I think the ONS have made a mistake by failing to strongly disown the
statistical adjustments made after the 91 census and admitting that the
estimation process for the MYE of population have been very wrong in some
cases. In the case of the latter, they should take immediate steps to get
new appropriate data sources in place. The delay in publishing revised mid
year estimates of population is unfortunate. Only then can we be expected
to believe the estimation process used this time round.

Simon




**********************************************************************************
This email may contain confidential material. If you were not an
intended recipient, please notify the sender and delete all copies.
We may monitor email to and from our network.

******************************************************
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]
*******************************************************