Opinion on an Opinion Poll

I've been busy in the outerweb (you see what I did there?) so this post is overdue but it was always coming. I've read a couple of other blogs on the subject so I'm confident that what I wanted to say is reasonable. It concerns the opinion poll carried out for the Press and Journal, my local newspaper. There are two articles which, I think, use the poll in a slightly confusing way. I don't think the P&J have been intentionally misleading but I do think that one or two points need to be clarified.

The articles are:
Labour set to hold key marginals - P&J Poll
and
Begg Faces Tough Re-Election Fight as 30% are Undecided
(Both from 28th April)

First a slightly trivial point - since Aberdeen South is one of the "key marginals" in question, these two articles do seem to be somewhat contradictory. A pedants quibble perhaps. My real concern is with the figures provided by the poll. The problem is illustrated by this quotation from the second article:

"Labour's Anne Begg faces a tough fight to be re-elected in Aberdeen South, where almost a third of voters have not decided whom they will back, according to a poll for the Press and Journal.
The poll shows that, despite Miss Begg's high profile, a general slump in support for the Government will lose her votes next Thursday. As many as 38% of those canvassed in the constituency said they had voted Labour in the 2001 general election, but only 21% said they would do so this time.
Just 9% of voters are sure they will vote Liberal Democrat, while at least 12% of the sample voted for the party in 2001."

12% of those polled said they had voted for the Liberal Democrats in 2001.
The BBC says that approximately 28% of people in Aberdeen South voted Lib Dem in 2001 so we're missing a large number of Lib Dem voters. Where are they? The Labour figure is correct, it's about 38% in both cases. The Lib Dem figure is hugely wrong, 16% of Lib Dem voters have disappeared. That's just plain spooky!

After reading this post from Stuart at Independence, I think I understand what has happened. It is just as well because I'm not very good at understanding the mechanics of opinion polls. I think the P&J has extrapolated the constituency poll result from a poll of the whole of Scotland.
That way madness lies people! It just isn't a very accurate way to do this sort of thing. As I understand it, each constituency would have to be weighted seperately before any useful conclusions could be drawn. Since we're missing so many LD voters, this clearly hasn't happened.

I am cheered enormously by this from Stuart's post:
"If any Liberal Democrats out there are reading this, you will be very pleased to hear that the P&J did an even bigger hatchet-job on the LibDems chances in Aberdeen South. Do the sums yourself, but whichever way you look at it, Vicki Harris will be the new MP for Aberdeen South, and Annabelle Ewing will return to Westminster as the MP for the new Ochil & South Perthshire constituency."

I'm not good at sums and I'm even worse at counting chickens but I've got my fingers crossed.
Honourable Fiend suggests in this post that the P&J poll might be "slightly whiffy". Maybe I should order one of Polly's nosepegs after all.
(Not so I can vote Labour obviously. That will never happen while Mr Blair leads the party, sorry Polly.)
Next post - a comment on all the "back door action" we've been getting from Mr Blair. Unsurprisingly, and as has been mentioned a number of times before, it is a sorry tale of untruths, scare tactics, and desperate deception.
(It'll also be tomorrow.)