This story is lacking the necessary details to do a proper analysis, but taking it at face value (i.e. 638 people were flagged up as possible liars and those were the 638 investigated*), 40% of the people flagged up were in fact losing out due to ignorance. This raises a few questions.

If Lambeth benefits staff were to investigate 600 of those who weren't flagged up, would they find a similar proportion missing out?

How many of the 377 who had their benefits reduced had made mistakes as innocent as the 261 who had their benefits increased?

Is the issue not in fact fraud but rather that the system is too complicated? Note that at least 13% of those involved in the study were missing out.

* I hope this assumption is wrong, because otherwise the "pilot project" is of zero scientific value. In order to establish whether the stress detection technology is any better than tossing a coin or die, it's necessary to obtain estimates for the false positive and false negative rates.

Peter Taylor, Valencia, Spain

Why not try this out in the city, on the traders who started this economic crisis? Because there would be the usual outcry of money going abroad. Far better those big-time pilferers be caught than trying this further extension of this sinister society on those who, in comparison, do far less damage to society than benefit claimants!

Peter K Day, Doncaster, UK/ Yorkshire

Could we use it on Gordon Brown please

Charles Daniels, Lady Lake, Florida