U-turn if you want to, but only under certain conditions

Back when I was ranting about designing a better political system, I promised a more general look at when it's good to be wrong. Any useful system has to be able to self-correct when it makes mistakes, or it's doomed to failure.

The current political climate of so many u-turns that we've certainly gone more than 360 degrees by now seems like a good opportunity to start to flesh this out a bit more. U-turns are traditionally politically undesirable, whereas in science changing your mind is also known as "thinking". Or at least, it often appears that way. Maybe in practise it's not always as neat as that. As the BBC note :

YouGov polls indicate more people viewed U-turns as a good thing (49%) - rather than a bad one (23%) - during the pandemic. Even before the coronavirus crisis, though, YouGov's polling suggests more people saw policy changes as good than bad. 

Chris Curtis, political research manager at YouGov, said: "In most cases U-turns by the government don't end up moving the polls. "If anything, YouGov data shows that people seem to see them as a positive sign that the government are willing to listen and change their mind if people complain or situations change."

(Without going into details about the poll, it's worth remembering that the public vote on the government as a whole, not individual policies - and the collective effect can be different from the sum of the parts. Hence the poll may well be misleading, which would explain why Labour are now on level pegging with the Tories.)

In response the scientist in me wants to shout, "hooray, the public understand that learning is a good thing !". On the other hand the political activist in me wants to fall into despair, wailing, "don't they understand that these people are dangerous authoritarian clowns ?". I strongly suspect that I'm both right.

To understand this, let's ask a question : what would the ideal u-turn be ? I suppose it would be if a decision was made in true accordance with the evidence, then reversed when new evidence showed the exact opposite. You couldn't really blame anyone for doing that. Politically, you'd also have to prepare people ahead of time, warning them about possible, uncertain policy changes. You'd have everyone prepared so that the change could be implemented as smoothly and as swiftly as possible. And your switch would be from one course of action that was bearable (neither harmful nor helpful) to one that was actually beneficial.

The extreme opposite would be a combination of first following the evidence for a beneficial result, then switching to an actively harmful policy in spite of the evidence, while continuously insisting ahead of time that you wouldn't change course and so springing it on everyone at the last minute. And a persistent climate of uncertainty is damaging in itself : people lose faith in the government's ability to do the right thing or ensure that it will help them when it decides to change tack. More directly, an inability to plan ahead causes obvious problems at all levels.

From this we may conclude that the important aspects of a u-turn are :

• How the change corresponds to the changing evidence
• Whether the change is positive or negative in terms of its effects
• How well you prepare people for the change ahead of time.

The current barrage of u-turns is full of aspects of all of these, good and bad. A current list can be found here, though unfortunately the once-proud Independent is now so rife with adverts that the text is barely legiglibe. Let me reduce this list to its basics :

1. Right to remain in the UK for foreign NHS staff, initially not allowing this for anyone except doctors and nurses but then expanding it to all care workers
2. Removing the NHS £400 surcharge for non-EU migrants after insisting they would have to pay
3. Allowing MPs to vote remotely after insisting they would have to appear in person
4. Giving free school meals throughout the holidays after saying the scheme would end
5. Declaring that the UK would have a "world beating" contract tracing app and then scrapping it altogether
6. The whole exam results fiasco, in which the government insisted on using a wealth-biased algorithm to compute results despite knowing this didn't work, and then switching to teacher assessment instead
7. Giving a short extension on an eviction ban shortly before it was due to end
8. Continuously changing advice on masks

Which I think is by no means a complete list, but it'll do. I'm not going to go through all of them anyway.

For the most part, the course changes do at least seem to have been for the better, so in that sense the public are right to approve of them. But rarely (if ever) have they been implemented in a sensible way. Worst of all these by far was the initial strategy of seeking herd immunity and then switching to total lockdown. This did go from bad to good, but it did not follow the evidence - the government appeared to have listened to a minor and bizarre sect and not the wider scientific community. It wasn't some obscure and esoteric piece of virology : it was blazing obvious that the mortality rate would be far too high for the herd immunity strategy to be endurable. Any warning of the change of strategy anyone had ahead of time came only from watching other countries. The switch was so fast that it cannot possibly have been due to following the evidence, which was there from the very beginning.

Basically this particular u-turn was a bit like walking towards a hail of machine-gun fire and then deciding, after having received a nasty nick on the shoulder, to run as fast as possible in the other direction. Yes, they did the right thing, but they should have been aware of that from the very start. That they took so long to realise this is a damning indictment of government incompetence, not a sign that they were responding to changing evidence.

A generally more positive example is that of masks. Again the switch was from bad to good, but this time, while the evidence certainly did shift only very slowly, the policy shift was still too slow. The country seemed to have real difficulty grasping the idea that masks are protecting everyone around the wearer, not the wearer themselves, while other countries not so very far away embraced this wholeheartedly. I mean, it's common sense : if you're being asked to cover your cough, then wearing a mask can only help. Fair enough that there was (and continues to be) some uncertainty on this one, especially currently with school mask policy, but it could have been implemented in a clearer, more decisive way.

While most of the changes were all but inevitable, what was a matter of pure choice was political insistence that they wouldn't happen. Treat NHS staff like dirt because of some trivial associated costs ? Changing course was a good thing; insisting that this wouldn't happen is what's problematic. Likewise, there was no need at all to insist on the high, superlative-laden rhetoric about the contact tracing app. The failure of the app is not the fault of ministers, but making promises they couldn't keep was absolutely and entirely their fault. Yes, they're now being slightly nicer to the foreign staff in the NHS, but only because their starting position was about as bad as possible; no, Boris Johnson didn't personally write the code for the tracing app, but there was no need to make exaggerated claims about its efficacy. Similarly for the exam algorithm : admitting it doesn't work is fine, but you should have been able to test this and spot problems well before using it in anger.

In short, when you start off by saying shitty things about your foreign workers that you never needed to say, or by ignoring obvious scientific evidence, or by throwing out undeliverable promises and unsubstantiated figures (a practise this government continues with gusto), then doing a u-turn doesn't much absolve your incompetence. Sure, it's way better than sticking to your guns and continuing to be an utter twat, but it doesn't mean you're listening. It doesn't mean you're learning. It means you're held hostage to events and your own stupidity. It means you'll make all the same mistakes next time around. U-turns are bad when a more competent administration would not find them necessary because they got things right from the start.

(As a side-note, I'll also point out that this is particularly true in the current case where the government has a thumping majority. In principle it could carry on regardless and outvote everyone else. That it doesn't means that it's clearly aware of its own stupidity, or at least unpopularity.)

This has gone in a more political direction than I intended, so what does this have to do with science ? How come science doesn't have these same problems with changing its mind ?

In the past I've suggested that in part it's because scientists don't have to take nearly so much flak from the media. And it's true that politicians get raked over the coals for literally everything they say, at least by someone, while scientists are treated very differently. But, while perpetual criticism is definitely a problem, this comparison to science is too simplistic.

In fact, science as a whole doesn't actually do that many u-turns. Rather, the emergent consensus arises through changing probabilities on a whole array of theories. That is, there are usually a multitude of different ways to explain any phenomenon, and the evidence weights each differently. As the evidence changes, different explanations change in plausibility, only on very rare occasions going from outright rejected to completely accepted (or vice-versa). Far more often, things just become more or less likely. Science has discoveries and incompleteness more than it does naked contradictions.

This is clearly quite different to a political u-turn, in which someone has to stand up and say, "I was wrong about that" and take the heat for it (or more often try to deflect the heat with a shield forged from purest bullshit). In politics the change is specific, away from something previously endorsed as beneficial to one previously denounced as unworkable or harmful. Science rarely has the direct equivalent, which would be to declare something to be impossible and then pronounce it as certain. It just keeps assessing probabilities, which is not the same thing.

And the emergent nature of the consensus is different from an individual having to confront a policy change. While individual scientists do make decisions and consciously change their opinion, such that any given moment one may find a plethora of contradictory views on a topic in a rapidly churning sea of information, the collective effort is slow, methodical, and careful. It's this emergent nature of the consensus which most distinguishes scientific findings from political judgements. Watch an individual scientist over time, or better yet pick a series of individuals all at once, and what you'll get is not so different from a bunch of squabbling good-for-nothing politicians. For politicians, on the other hand, individuals are far more important, as it's often at that level - not at the consensus level - that arbitrary decisions have to be made.

Nobody has to do that in science - or at least, not in the usual sort-of scientific discovery process. Nobody gets to say, "this is true now". Since no one individual is usually responsible for a major breakthrough, it's vanishingly rare that the media need to single them out for bad behaviour or silly mistakes - simply because no individual was responsible for the emergent viewpoint. Sure, they have changing views on their own particular little corner of reality, but usually these are so small that they're only important as part of the collective whole.

And scientists have only opinions and beliefs. They don't have authority or set policy. They determine the facts and theories on which politicians have to act, but they don't have to take those decisions themselves. They can and do advise on a course of action, but not actually make the decision. It'd be a bit weird to pour scorn on someone for changing their own mind, which is not the case when it comes to that person taking action that impacts others. Science is a form of exploration, and consequently expects and desires change, knowing ahead of time that it will learn from its mistakes... whereas politics is expected to set stable policy that everyone can live by. Both the purpose and standards of the two systems have some marked differences.

So there we have it : a set of criteria on how to judge if a political u-turn indicates competence or ineptitude, and some possible reasons why this isn't an issue for science. Whether this actually helps in developing a better political system I leave to a future post.