The Death Of Expertise

But social scientists know that. The astrophysicist, and the plumber, have many hard rules, and because of this, there are many situations where their training will force them to shout "No! Stop!" at something you're doing.

Social scientists, by-and-large, don't have very many hard rules... and so they don't often shout "No! Stop!"

They are aware, far more than the members of the hard sciences (who are used to most things they deal with being either hard rules, or debunked nonsense), exactly how vaguely-implied or well-proven each of their theorems are; how controversial or universal each stance is. And because of this, when a social scientist is tapped on practical matters of public policy, they will more than anything be hedging, not strengthening, the prior stances held by the policy-makers, by reminding them of all the things we can't assume and aren't sure will work.

(On the other hand, when social scientists are tapped by journalists, or when they write twee little NYT bestsellers, the result is more an exposure of the scientist's personal position in debates internal to the academic sphere. This has nothing to do with what a social scientist would give you if you paid them to do social science for you.)

The most useful thing a person with training in the social sciences can tell you, of which the general population isn't aware, is what you should be less sure of. No matter where a social scientist falls in matters of their own field's debates, they'll know which things are debated, and therefore can shout "No! Stop!" when you're about to base a policy decision on, for example, "the proven effect of Keynsian stimulus." (This is closely related, I think, to being trained in impossibility results. It takes you a computer scientist to tell you that, no, your new database can't be C, A, and P.)

Looking at internet threads engaging in amateur sociology is a decent demonstration of why sociological expertise is an actual field of expertise. It's not my field, but I've at least read a 101-level methodology textbook, which I think would improve the state of things considerably if even that level of knowledge were more widespread. Doesn't require a degree, but does require a willingness to learn, whereas I think many people with an anti-expertise viewpoint are more interested in rationalizing why they shouldn't have to learn it (the "only engineering expertise is expertise" view is one variant).

Computer programming is also greatly aided by the existence of machine-checked proofs: the ultimate impartial mediator in any dispute between self-styled experts. Unfortunately, such a convenience does not exist outside a handful of fields. What we're left with, then, is a credentials-measuring contest.

He mentions that people are not equipped to judge evidence properly, which is often true. However, people are also not well equipped to judge expertise either.

In the past few decades, we have witnessed what appears to be massive failure of experts on a number of fronts. In the media, people who certainly look like experts turn out to be con men... when well dressed "doctors" and "scientists" lie and trick people into making terrible choices, we grow suspicious. Anyone can pretend to know what they are doing, and blind trust of such people is a dangerous path to follow. Even the antivax people have what appears to a layman to be experts backing their side of the story.

Further, specifically in the areas of social policy, finance, and politics, the past decade or more seems to many to have been a disaster. Thousands of experts missed the financial disaster, and, from what the story looks like to many, bungled the recovery. The news that sells is all about disaster and the unexpected; things that experts failed to predict.

In is unsurprising, given these circumstances, that the title of expert is viewed not with blind respect, but with suspicion. If you claim to be an expert, then you are instantly placed in the same category as people who just want to exploit me for a quick buck.

I don't have a solution, but it is an explanation, and "trusting experts more" is not going to happen.

Sort of off topic, but putting developers in with doctors and engineers is not really legit. The latter two have rigorous tests and certification standards, while developers (myself included) have nothing of the sort; furthermore, much of the methods and such used by developers are not founded at all in research or a scientific method, and are usually of an even lower standard then most soft sciences. (They at least have peer review, we mostly just have blog posts and books from random people from industry).