Henry Dimbleby always enjoyed his lunch, as I can testify from the days, perhaps 20 years ago, when we used occasionally to eat together in The Daily Telegraph's canteen, which in those days was a boat moored next to South Quay, in London's Isle of Dogs.

His physical resemblance to both his father David and his grandfather Richard – the Almighty, when designing the Dimbleby clan, clearly decided not to stint on materials – was a further clue that dinner was often uppermost in his mind. Confirmation came when, despite his glittering journalistic heritage and his own evident talent for the trade, he jacked it in to set up a restaurant called Leon.

Its model, according to Mr Dimbleby and his business partner John Vincent, was a vision of "what McDonald's would be like in heaven". In other hands, it's easy to imagine how repulsive food based on this right-on agenda might have been, but actually Leon produces fairly interesting, healthy, tasty, fast-ish food which, if not exactly cheap, is reasonable value. As a result, this enterprise is now a highly successful chain and when Michael Gove, the Education Secretary, chose to look at school meals, he called them in to produce a report, which appeared last week.

Its conclusion was that packed lunches should be banned, and universal free schools meals introduced in England, which would, of course, be carefully costed and nutritionally balanced. The Scottish Government was quick to announce that packed lunches won't be banned north of the Border, presumably because it thinks it can recognise an idiotic and potentially unpopular idea when it sees one; and I think we can say with confidence, because politicians south of the Border suffer from similar delusions, that it won't happen in England either. Just yet.

But watch this space. The Westminster Government has also just announced that it will not be proceeding with plans to ensure all tobacco products are sold in plain packaging, while Holyrood is pressing ahead – just as it is with the minimum pricing of alcohol, which (although the Coalition previously expressed its enthusiasm for it) was shelved in England and Wales.

In each of these cases, a public health issue is being presented and a desirable outcome assumed. The arguments which follow are overwhelmingly about whether the particular policy recommendations being put forward are founded on evidence-based studies of their possible efficacy, or about secondary considerations, such as whether the minimum pricing of alcohol would breach EU law (which most experts think it would), or whether tobacco tax revenues are higher than the cost to the NHS of smoking-related conditions (which they are, by some distance).

As a result, we are inclined to get tied up by the problems and the hope that they can be resolved. No-one disputes that morbid obesity, as opposed to the slightly elevated health risks of being slightly overweight, is becoming a major problem with serious consequences for health spending. No-one disputes that smoking is dangerous, even if they think the evidence that passive smoking is a major health risk is flaky. No-one disputes the dangers of serious alcohol abuse, even though the evidence is that moderate drinking is healthier than total abstinence.

And because these are real problems, with real costs in personal, social and financial terms which suggest the desirability of nudging behaviour in fixed directions, the scientific evidence is very often presented in highly dubious forms.

Attempts to ban e-cigarettes, or snus (a chewing tobacco popular in Sweden) are obvious examples of scientists and policy makers ignoring the reality – which is that they are far preferable to the real dangers of smoking – because of a desire to eradicate tobacco altogether. Similarly, out of many studies, there is only one (conducted at Sheffield, and widely criticised for its methodology), which suggests that higher pricing would reduce alcohol consumption among the heaviest drinkers. Despite my admiration for Mr Dimbleby's restaurants and their ability to produce a healthy lunch, I can't quite see that his report's assertions about the unhealthiness of most packed lunches, compared with the average school dinner, would bear the closest analytic scrutiny, either.

Though I have no difficulty in accepting the scientific evidence that we are in a period of – perhaps very significant – climate change, and that some of it is probably caused by human activity, that doesn't obscure the fact that the policy recommendations based on this evidence are often informed by consensus about shaping public moods and perceptions rather than the raw data.

The evidenced-based argument – whether accurate, distorted (but with the best of intentions), or downright idiotic – is still a red herring unless and until you accept one premise: that the Government has a right to dictate how you live your life. For example, there's quite good evidence that people who walk two or three miles a day two or three days a week live longer, and are happier and healthier than those who don't – even when you ignore their other lifestyle choices, such as diet, drinking habits and smoking.

So the obvious question is when governments are going to make this kind of low-level aerobic exercise compulsory, as it was, you'll remember, in George Orwell's Nineteen Eighty-Four. What's that? No government would dream of doing such a thing?

Well, actually, Justice Antonin Scalia of the US Supreme Court has already given his opinion (admittedly as a contrarian position) that failing to exercise undoubtedly "causes healthcare costs to go up", and that any state-provided health system should insist on imposing exercise on those who benefit from it.

It is vital that the putative benefits of public health measures be evidence-based, and that our understandable desire to tackle obesity, or smoking, or alcohol abuse should not be swayed by shoddy claims that these issues can be magically disposed of by some new restrictive measure. But much more important is the question of whether the restrictive measure can be justified, even if it would lead to the improvements which are presented by its advocates.

If you could prove conclusively (and good luck with that) that obesity could be eradicated by banning chips, I'd vote to keep chips. If we would all live two years longer, and cut the NHS bill by 10%, by being forced to walk for three miles, three times a week, it would still be wrong.

If you want to ban alcohol, or tobacco, by all means try that: the history of Prohibition in the US may give you an evidence-based clue about what you're in for. But don't pretend that, at root, what you want to do for our good is any different from the prescription offered by Fascism and Stalinism.