Daniel Hannan recently tweeted:

Around 80% of us say we support the lockdown. But, looking around me, I'd say that no more than 20% are still observing it rigorously. Is this a case of what economists call "revealed preference" - or do people want everyone else to apply stricter rules than they do themselves?

It might well be neither. Instead, it's an example of the sampling bias. The people we see are, by definition, those who are outdoors and thus who are disproportionately likely to be breaching the lockdown. What Mr Hannan isn't seeing are the countless thousands of us staying indoors and observing the lockdown. He's failing to appreciate that what he sees is a biased sample of what there is.

In fairness to him, this is a common error, as experiments by Benjamin Enke have demonstrated. Here are a few other disparate examples of a similar mistake.

Benefit fraud. People greatly over-estimate the extent of this. A survey by the TUC a few years ago found that "people think that 27 per cent of the welfare budget is claimed fraudulently, while the government's own figure is 0.7 per cent." One reason they do this is that they can see fraudsters whereas they don't see the millions trapped indoors by disability or depression.

Opportunity cost. As Frederic Bastiat pointed out, we see the shopkeeper spending money to replace his broken window but don't see that he would have spent money elsewhere had his window not been broken. We thus wrongly infer that some activities - be it breaking windows or going to war - boost economic activity when in fact they often don't.

The unseen counterfactual. Because we don't see the road not taken, we don't benefit from the Jim Bowen effect: "here's what you could have won." We don't therefore blame governments sufficiently for policies which impoverish us relative to a plausible counterfactual - such as austerity or Brexit.

Social transmission bias. In a recent paper David Hirshleifer describes how people's savings and investment decisions are distorted by the biased signals they get from others. Because people are more likely to discuss their successes than their failures, they talk more about low-return but high-risk strategies that happened to get lucky. This leads listeners to believe that stock-picking works better than tracker funds and that speculative shares pay off better than defensive ones when in fact (pdf) they generally don't.

Survivorship bias. The shares that are listed on the stock market are by definition those that have not gone bust. They are therefore a biased sample of all stocks. Because investors don't fully appreciate this bias, they under-estimate the chances of corporate failure, especially over the long-run, and so over-estimate their chances of picking good stocks. In fact, as Hendrik Bessembinder shows, most stocks actually lose money over their lifetime, and stock market rises are due to great performance by a tiny minority of shares.

Biased news. As the ancient saying goes, "'dog bites man' is not a story but 'man bites dog' is". What we see reported on the TV are, therefore, unusual events. But by definition these do not describe reality. It's partly for this reason that the fear of crime has not fallen as much as actual crime has. Also, in reporting salient events, the news understates slow-moving but world-changing trends - be it the decline in global poverty, stagnation in productivity or long-term decline in government borrowing costs. In this sense, the news - even if presented wholly honestly - distorts our perceptions of reality.

Immigration. People see what they think to be an adverse effect of immigration: "some immigrants arrived a few years ago and I've lost my job since." What they don't see, however, are the many mechanisms which counteract this apparent effect.

And herein lies the point. The social sciences are, as Jon Elster said, fundamentally a collection of mechanisms. But many of these are unseen. It is the role of social science to expose these mechanisms, and to show us that what we see is not all there is. As Marx said: "If there were no difference between essence and appearance, there would be no need for science." Or as Marvin Gaye, another great social thinker,
famously put it: "believe half of what you see and none of what you hear." Mr Hannan could learn something from him.