* Weak official GDP data at odds with upbeat surveys
* Conflicting signals could heighten tensions on MPC
* QE seen likely to expand but economists wary
* BoE may not take Q3 contraction at face value
By Christina Fincher
LONDON, Nov 3 (Reuters) - When Bank of England policymakers meet this week to discuss whether to pump more money into the economy they will have to piece together some contradictory data.
Official figures show Britain's economy shrank for a sixth consecutive quarter between July and September, making the current recession the longest since the Second World War.
Private surveys, on the other hand, suggest the economy has not just stabilised but is expanding. Retail sales are growing at their fastest rate since December 2007, consumer confidence is on the up and even manufacturing is expanding at its fastest pace in two years.
This dichotomy poses a headache for BoE watchers. They must not only make a judgment on the true state of Britain's economy but also on which pieces of data the central bank is likely to prioritise.
Peter Westaway, chief European economist at Nomura, reckons a stampede to change forecasts following the shock contraction in UK third-quarter GDP is misplaced.
"I don't think the BoE will want to put too much weight on any one piece of data," said Westaway, who has just joined Nomura after 15 years at the central bank.
"It may well be a split decision and a close call, but overall I still think there is a strong enough case to pause."
HERD MENTALITY
Economists have had a poor year for second-guessing Bank of England policy moves and have chopped and changed in recent months.
In July most thought the BoE would extend its asset purchase scheme. It didn't. In August economists were split on whether the BoE would pause in its programme or increase it by 25 billion pounds. In the event, it opted for a 50 billion pound expansion, justifying the move on a surprisingly sharp drop in second quarter GDP.
The BoE's focus on the "output gap" -- the spare capacity in the economy that must be used up before inflation can take hold -- helps to explain why preliminary third-quarter GDP data on Oct. 23 sent such shock-waves through financial markets.
The figures showed Britain's economy did not pull out of recession in the third quarter, as widely expected, but instead contracted by 0.4 percent. Within days, half the economists who had previously predicted the BoE would halt its QE injections in November switched sides.
The latest Reuters poll shows more than two-thirds of economists now expect a further QE boost this month. But, after so many wrong calls, conviction may be lower than this figure suggests.
"Expectations have tilted in favour of a QE extension but no one wants to put a lot of risk on the table," said Simon Hayes, UK economist at Barclays.
Peter Dixon at Commerzbank joins Barclays in forecasting a 50 billion pound QE boost this week but assigns the probability to this outcome at "60 percent at best".
QUESTION OF TRUST
Despite the impact of the data, it is far from certain the BoE will take the reported decline in third-quarter GDP at face value. In the past, the central bank has chosen to substitute or "backcast" its own data into its models when it has doubted the accuracy of the official figures. It could so do again, particularly if it finds it hard to square with evidence from other sources.
Britain's third-quarter GDP contraction did not just raise eyebrows in City dealing rooms. The BoE's own forecasts, published in August, had assumed growth of around 0.2 percent between June and September.
Preliminary GDP figures are based on only 40 percent of the data that eventually feeds into the final numbers and are notoriously prone to revision.
George Buckley, chief UK economist at Deutsche Bank, calculates that in the decade to end-2007 the correlation between the preliminary and the final estimate of GDP was only 10 percent.
Revisions can be dramatic. In the fourth quarter of 1998, the preliminary reading put growth at 0.2 percent, but the final reading showed it at 1.1 percent.
Sample differences and the fact that surveys tend to ask qualitative rather than quantitative questions could all have contributed to the different results.
But even allowing for these factors, the gap is hard to close.
"We're getting into the realms of forecasting what the Bank of England will be backcasting. It makes the outcome that much harder to predict," said Buckley. (Editing by Stephen Nisbet) (uk.economics@reuters.com +44 207 542 7748; Reuters Messaging christina.fincher.reuters.com@reuters.net))