Here are three questions and three predictions.

Question 1 Should Britain move towards a low-tax, small-government society? 2 Should the private sector play a much bigger role in the National Health Service? 3 Should there be far more redistribution from rich to poor?

Prediction 1 Most New Statesman readers will respond no, no, yes. 2 The minority of Conservative readers who read the NS, to keep an eye on what the other side thinks, will respond yes, yes, no. 3 What both groups of readers will share is the passion of their answers.

My point here is not to debate these policies, but to observe one of the differences between normal voters and political junkies. Understanding this difference can hold the key to winning elections; and special YouGov research for the New Statesman shows how.

The difference can be illustrated by one of the questions. Suppose you feel that strongly about the role of the private sector in the NHS, either for or against. That is a positional view. But suppose you don’t mind that much either way, and all you want is prompt, high-quality care when you need it. In that case, yours is a valence view.

Most politicians, activists and commentators are full of positional views. But millions of swing voters aren’t: they take a valence view of politics. They judge parties and politicians not on their manifestos but on their character. Are they competent? Honest? Strong in a crisis? Likely to keep their promises?

Branded for life

To explore this, YouGov took a series of political controversies and asked people which of four options they favoured – in each case, we offered three positional options (“left”, “right”, status quo) and one valence option. You will find the full wording and survey results summarised online; a table covers the positional choices we offered to respondents and spells out the valence option (log on to  newstatesman.com or visit yougov.co.uk).

At first sight, the news for progressives is bleak. The “left” view lags way behind the most popular option on all six questions (see below). That is not all. Look separately at the social and “nationalist” issues and you find that the valence option is by far the most popular on all three social issues, while the right-wing view is the most popular on all three “nationalist” issues.

Now, many NS readers will find this hard to accept. These findings just don’t square with the way they – we – and most of our friends discuss politics. The trouble is, we’re not typical. Take the issue of  redistribution. Among those who identify “very strongly” with Labour, 46 per cent want the government to do more to help the poor, while 14 per cent think the poor should take more responsibility for their own plight. But among those with no allegiance to either party, just 9 per cent back redistribution, while 29 per cent want lower benefits. In this group, the valence option is by far the most popular. Forty-eight per cent think that what matters most is not the size of the benefits bill, but how fairly benefits are distributed.

And here’s the killer fact: very strong Labour identifiers comprise just 5 per cent of the electorate. Those with no allegiance comprise 24 per cent. These are the floating voters who decide elections.

In short, Labour can’t win the votes that matter simply by promoting progressive policies; it must win the valence war. This does not mean ignoring the task of policy formation. Far from it. The task is to burnish the Labour “brand”, which is fundamentally similar to building a commercial brand. BMW has a strong brand. This is not because many BMW owners know (or care) about, say, their car’s electronic software or the composition of its engine casing. It’s because owners trust BMW to get these things right. And because BMW’s engineers and production processes have proved reliable down the years, so the brand has grown strong. Likewise with political parties. The “engineering” – policy formation – is vital; but it will be electorally effective only in so far as a party’s policies, collectively, enhance its overall brand image.

Rewards for respect

This analysis can help Labour even on the “nationalist” agenda. Forget trying to persuade floating voters to like Europe, immigration or shorter prison sentences. Those arguments are unwinnable, at least in the short term. What can win votes, or at least avoid terrible losses, is evidence that the party will handle these matters honestly and competently.

The most dramatic example of valence politics trumping a populist positional stance occurred in the Romsey by-election 12 years ago. The news was dominated by reports of asylum-seekers escaping from the Sangatte camp near Calais and crossing the Channel. The Tories sought to defend one of their safest seats by mobilising public anger over the issue. The Lib Dems seemed to be on a hiding to nothing –yet won the seat on a huge swing. In the event, the Tories lost votes because they appeared to be exploiting the issue cynically; the Lib Dems, who avoided crude populism on this occasion, were seen as more principled. The Tories’ positional view chimed with most Romsey voters, but they lost the valence war.

By the same token, David Cameron could lose votes by promising a referendum on the EU, even though most voters want one, if he appears to be buffeted by events and mesmerised by the threat from Ukip, rather than taking a lead. In the end, the voters who decide elections judge the parties and their leaders by their character. Parties win valence wars not when they abandon unpopular policies to appease the public mood, but when they show by their behaviour and the quality of their leadership that they deserve respect. Thatcherism in the early 1980s and New Labour at its most popular were powerful examples of strong valence brands. The valence winner of 2015 has yet to emerge.

Peter Kellner is the president of YouGov