James Frayne is Director of communications agency Public First and author of Meet the People, a guide to moving public opinion. The focus of this column is Theresa May’s conservatism for “ordinary working people”.

There’s a growing trend to say policy decisions must be data-driven. We hear endlessly about the need for “evidence-based” policy making and there’s an ever-increasing role for so-called policy nerds within political organisations and the media. This is generally welcome: politicians and advisers must make decisions based on facts.

With such a trend, you’d expect the public to become happier with the direction of government policy over time. Well-evidenced and targeted policies should produce better outcomes and more reasoned public debate. That’s not how things are going. If anything, the signs are that people are becoming less happy with the big decisions politicians take.

Brexit is just the most obvious recent example where the public have rejected politicians’ and advisers’ arguments – and their use of supposedly unarguable evidence – that the country and those that live within it would better off.

Why does this happen? One of the problems with such a self-consciously evidence-based approach is that it assumes three things: (a) that people should always be able to see and appreciate the bigger picture; (b) that the perceived greater good for the country equals the greater good for communities; and (c) that there are no difficult trade-offs in policy and evidence is always clear.

There’s a particular question I’ve begun to ask more in meetings, partly because it’s important, partly for effect: “what would we do if we were making decisions purely on behalf of our own families?” This isn’t about making policy by anecdote or personal experience, but to ask ourselves whether we’re in touch with ordinary voters.

Nowhere is this more important than in the field of education policy. Many people in the political world oppose more choice for parents in the form of grammar schools and greater private provision – while sending their own children to such schools. Similarly, many support an extension of non-academic courses that they’d never consider for their own children.

This is not about what I think is the right education policy. But asking what you’d do as a parent is not an irrelevance. It’s an important way of making policy that reflects public concern. We should be extremely wary of advocating a particular policy that we’d personally ignore for our own families.

What goes for education policy goes for other areas too. Would we be happy to see the person that burgled our house or our friend’s house be given another slap on the wrist rather than a custodial sentence? Would we honestly be happy with a new hostel for the homeless on our quiet street? Would we accept seeing our parents in a badly managed care home? Do we personally want to pay higher prices for alcoholic drinks because some abuse alcohol?

For many people, the answer to each of these will be “no”.

Policy by anecdote is rightly criticised. And to the extent that there’s a choice to be made between the use of evidence or opinion, that’s no contest. Evidence and data should win out. But when people consistently behave or feel differently than the policies they support for others, it’s a bad sign. We need a personal sense-check for every proposed policy.