You might have thought after the knife crime fiasco that No 10 Downing Street would have realised the danger of “mix and match” statistics.

Not so. The Prime Minister has been pulled up by the UK Statistics Authority for making claims on a podcast about net immigration that used data from two data sets. He compared Long-Term International Migration figures for 2007 and 2008 with figures for 2009 that came from the International Passenger Survey.
 
These figures showed, Mr Brown said, that net inward migration had come down from 237,000 in 2007 to 163,000 in 2008 and to provisional figures of 147,000 in 2009 – a rapid fall.
 
For consistency, he should have compared provisional figures throughout. Since the latest figures run up to June 2009, that means he should have used the provisional figures published for June 2007 and June 2008. The comparison would then have been June 2007: 170,000, June 2008: 168,000, and June 2009: 147,000.
 
It would have made the same point, but using consistent data. But it would not have implied quite such a sharp fall, and his point would have been weakened.
 
The full data are spelled out in a note from Sir Michael Scholar to the Prime Minister which recognises that in a speech today Mr Brown had used the correct  figures. That implies that he had corrected himself before Sir Michael’s letter landed, a fact used with advantage on Channel 4 News last night by migration minister Phil Woolas in his defence of the Prime Minister.
 
It would be interesting to know if this correction genuinely came unprompted, or was made after No 10 had been alerted by somebody that an investigation was under way. Sorry to seem suspicious, but I would guess the latter is more likely.
 
The full letter and note prepared by ONS is available here on the UK Statistics Authority website. The table is copied below: IPS stands for the International Passenger Survey, LTIM for long-term international migration figures.
 
                  
It can be seen that there is generally a correction between the IPS and LTIM figures, accounted for by migration to and from Northern Ireland, asylum seekers not counted by IPS, and switchers who change their plans once here about how long they plan to stay. These typically increase the IPS figures by anything between about 15,000 to 40,000. Last year it was 34,000.
 
Mr Brown’s figure for 2009, which he gave as 147,000, might therefore lie between 160,000 and 185,000. It could in theory be greater, and not smaller, than the 2008 figure of 163,000. But we won’t know until the election is long over.
 
One final mistake by No 10. The 2007 figure, quoted by the PM as 237,000, was in fact 233,000 after revision by the ONS in November last year. Makes your head buzz, doesn’t  it? But nothing like as loudly as somebody's is buzzing in No 10.