Don't Forget the Silent Majority

The Times made it the morning's main story: "Half of all servicemen say they want to quit". Research by the Ministry of defense involving more than 24,000 military personnel, had discovered, the paper explained, that 47% of soldiers and army officers thought regularly of resigning, with 44% in the RAF doing the same.

As for morale - listed with equipment and pay as the chief source of discontent - 59% in the army thought it was low, and that figure rose to 72% in the RAF. The Guardian was less excited, putting the story at the foot of page seven, but the essence of its report was the same: 47% of potential defectors in the army, 44% in the RAF. These figures will from now on constantly recur in parliamentary debates, party political speeches and all over the media. So perhaps it's worth pointing out that their accuracy is doubtful, and the margin of error involved is beyond calculation.What these reports did not say, though the BBC did, was that of the 24,000 "involved", some 9,000 responded. This might sound like a hefty sample - it's nine times what most political pollsters use - but having a very large sample only helps if that sample is reliably representative, and there's absolutely no reason to think that this one was. It is usual on these occasions for those who are discontented to seize their chance to air their views, while those who are less unhappy cannot be bothered to answer.

So when the Times and the Guardian say that 59% of army personnel questioned rated morale as low or very low, what we're really being told is that almost 20% of those the MoD set out to survey gave those answers; not quite such a daunting figure, perhaps. That certainly doesn't mean that it isn't true that many in the armed forces are unhappy with morale and equipment and pay. What is wrong is presenting the figures that appeared this morning as if they were the indisputable truth, scientifically assessed. All surveys, all polls, should be published with some indication of the numbers of those involved whom the pollsters class as don't knows and won't says. Show the troubled minority certainly, but don't in doing so forget the existence of the silent majority, which here outnumbered respondents by a margin of 5 to 3.