Comment by jlokier | Hacker News Reader

Comment by jlokier | Hacker News Reader

Parent Full thread

> What intelligent person would trust institutions that are inconsistent [...]

An intelligent person would trust institutions that change their mind more, because it's the right thing to do.

Institutions that remain utterly consistent in the face of evolving knowledge are the ones you should be wary of.

I saw an analogous thing in the much simpler world of signal processing and statistical estimation. Estimators and control systems sometimes oscillate as new data is acquired. It seems counterintuitive, more data should just improve accuracy, right? As if to converge toward an underlying true value? No, sometimes the best possible estimate and control output oscillates gently as more data is acquired, without any inconsistency. It took me a while to appreciate that.

Back to the big picture of institutions. Inconsistency over time may seem dissonant, but it shouldn't. It can be the most correct and accurate recommendation over time as new data and knowledge is accumulated. In an evolving situation, you should be seeing this.

The fact is people do cite inconsistency as a reason to disbelieve, and it puts public health policymakers in a dilemma. If they tell the truth and give the best advice to follow, people don't believe it because the truth is complicated and counterintuitive, and best advice rightly changes over time, and in different locations, circumstances, etc. So they have to walk a line between fully detailed truth, and simplified advice that people en masse are more inclined to believe and follow.

(I don't disagree with your other assertions about accountability etc).

> it puts public health policymakers in a dilemma

> they have to walk a line between fully detailed truth, and simplified advice that people en masse are more inclined to believe and follow

I know they are acting in good faith. However, the field also has a lot of received wisdom that may or may not be true. There’s clearly a bag of tricks they believe must be used to achieve a set of self-set goals, and I think a more straightforward approach along with dialogue would have been wiser in this instance.

The alternative is that some people will lose trust and won’t comply. You don’t want that if your plan requires everyone to comply.

> The alternative is that some people will lose trust and won’t comply. You don’t want that if your plan requires everyone to comply.

The dilemma they face is no matter what they do, some people will lose trust and won’t comply.

Getting an optimal outcome in the middle of that dilemma is impossible to get right. Not just difficult, but impossible due to the back and forth inter-reactions between individuals, groups and policy.

So the process inevitably involves some mixture of politics and data, and multiple points of view that differ. There's no getting away from it.

I'm not going to say they have got it right. The different points of view and debate have to continue after all, to pursue the "best" outcome whatever it may be.

Only that I have some sympathy with the public health messaging dilemma, and I hope other people can take something useful from the knowledge that public health policy people face this dilemma, and are not just incompetent, blunt instruments, or trying to hoodwink everyone.

Endless message mapping and behavior change campaigning was never going to cut it here for an entire pandemic. People know they’re being led by the nose.

I get it, it’s hard, which is why we give the CDC a ton of money every year.

I just don’t have much sympathy for people who deploy preplanned psychological tricks and then don’t get what they want. We deserve better, seriously.

> The dilemma they face is no matter what they do, some people will lose trust and won’t comply.

This is speculation. There are an infinite set of approaches for complex situations, that we have not found an approach that is consistently effective in no way proves that "no matter what they do, some people will lose trust and won’t comply", it only proves that they won't trust & comply under the approaches we have tested.

Such things seem to be easy to conceptualize when we are performing computer system analysis, but it seems like when we are embedded within the system being analyzed, we lose this ability.

> So the process inevitably involves some mixture of politics and data, and multiple points of view that differ. There's no getting away from it.

This seems true. But it seems to me that when making strategic decisions about how to go about this, we make plausibly unnecessary mistakes on a regular basis. One problem might be though: if we start discussing some topics in an objective, truthful manner (more facts & logic oriented, less "truthy" narrative oriented), the public might expect us to start discussing all things in this manner.

> public health policy people face this dilemma, and are not just incompetent, blunt instruments, or trying to hoodwink everyone

If the government, media, and health professionals would explicitly acknowledge the fallibility and errors of institutions, perhaps people would have more trust. But instead of describing reality as it is, we seem determined to stick with the simplistic fantasy land approach. This approach was extremely effective prior to the internet, but it seems insufficient now (except sometimes, like politics - here it continues to work like a charm, albeit with fairly disastrous outcomes).

Reply on news.ycombinator.com