Dismissal Of Subject Experts

Moved from SubjectMatterExperts as it is the main topic there.


Dismissal of Subject Experts

Many conservatives reject conclusions drawn by subject experts, such as biologists, climate experts, and economists. Asking them about it, I've concluded that for the most part it's not that conservatives consider them "dumb", but merely likely to be biased by political and/or personal beliefs such that the influence of bias overwhelms objective judgement.

I suspect a lot of this is merely projection because they know deep down that their judgement is influenced by personal and/or political bias. Conservatives often value "gut feelings" more than what most of us call objective science and logic. It may relate to their religious belief that God is guiding them in judgement, and God is going to be more reliable than "mortal science" because God is allegedly omniscience. They have a "special cable" to God's wisdom while "communist" subject experts don't. Thus, having a "rational" discussion with them based on analysis, details, evidence, models, and logic is often moot because their gut feelings have special powers over mine because they jack off less, pray more, go to church more often, or something like that.

Note that I am not claiming that all conservatives or Christians are like this, just the more stubborn ones.

I've experienced similar attitudes regarding IT. Such managers are so used to relying mostly on their gut feelings outside of IT that their logic in general atrophies. Thus, their problem behavior is not limited to just political and religious issues; it leaks into their IT judgement.

--top

I must admit that I have become more mindful of their skepticism that researchers will do or find whatever makes their wallet thicker or accolade board larger. If you are more likely to get a grant for concluding B than A, human nature is such that there may indeed be an inclination to favor B. Sure, there are checks and balances in science, but a lot of details that fly under the radar of these balances may tilt toward money and power. I'm not claiming there is a vast conspiracy, only saying bias may be playing a part and this issue is under-explored. Conservatives exaggerate the influence of bias on final results, but that doesn't mean bias is not there. -t

There are certainly researchers of that ilk. Within the research community -- where, by in large, its members are dedicated to science and extending human knowledge and capability -- we consider such "researchers" (I use the term loosely) to be scum.

Keep in mind that one may not be aware of their own biases. There are those who intentionally play games for money, but there may also be unconscious bias from those who believe themselves to be honest and open-minded.

I suppose, but it's pretty hard to unconsciously write up funding bids.

There are numerous ways to have bias sneak in. For example, which mysteries you choose to investigate (or request funding for) may be affected by bias. Some experiments you may work extra hard to get lots of sample points and some you may merely coast.

Sure. That's why there's an Ethical Committee that reviews funding bids and research methodologies, in order to help identify biases. Ideally, researchers wouldn't have to fight for funding and would have money pour onto them from the sky to pursue their personal interests, but, alas...

The Ethical Committee probably have their own biases, and if those line up with those of some researchers, it could be magnified. Bureaucracy itself is not sufficient to remove all bias. I agree an oversight committee is usually better than no oversight committee, but still no guarantee of bias removal.

Of course. Human biases are inevitable as long as there are processes that require human decisions. But what's the alternative?

I don't have one. The conservative "alternative" seems to be to wait until global warming is blatantly obvious such that there is no uncertainty about its existence, and THEN do something about it.

"Wait and see" appears to be fundamental to the "conservative" ethos, unless it's in regard to doing something about activity they consider morally questionable.

Some call it "folk science".

Some call "wait and see", "folk science"? That doesn't make sense.

It's sometimes waiting until something is plain and obvious before conclusions are made.

Ah. But that's not science, folk or otherwise. That's just asking for trouble.

It needs to be taught as a process, not as a social institution. The current arrangement is seen as a social institution by some, allegedly subject to massive GroupThink: self-reinforcement, and monetary enforcement of bias.


Top is condemning someone else for dismissing subject-matter experts? Interesting.

The "subject experts" on interpreters don't agree. (Plus, we don't have reliable surveys.) One paid implementer didn't tightly bundle the tag with the value "representation", for example. It smells like academic habit/tradition instead of absolute rules of logic or physics. WeAlwaysHaveDoneItThisWay? is not proof. -t

{Top, now really... Your so-called "paid implementer" had no credibility; his posts were such a muddle of technobabble and anti-mathematician rant that you can't tell what was actually bundled with what, only that as a "subject expert" he was a highly questionable source.}

"Muddle of technobabble", "questionable source"...Kettle? I saw zero evidence that it outright wouldn't function. Further, I borrowed the "internal variable name" concept from another commercial system; I did not invent it. You can perhaps argue it's a "bad design" per performance and/or maintenance effort, but it ran and sold. (I'd also argue it makes debugging easier by supplying a reference name for intermediate expression results. Clear ID's often make life easier for analysts/trouble-shooters.)

{"Clear IDs" may be a useful thing for a debugging tool to construct at debug time. They have no role in an abstract model of types, values, and variables in popular imperative programming languages. I'm sure there are implementations that used inefficient constructs or nonsense like "anonymous variables" or whatever. That doesn't legitimise their approaches; it only means they're badly implemented.}

That's your opinion, not an objectively-proven truth. Plus, having the engine use one method for regular runs and a completely different method for debugging mode complicates a system because it has to support two ways to do the same thing. Anyhow, this probably isn't the place to jump back into the "type fights".

{The "engine" already uses "one method for regular runs" and "a completely different method for debugging". The "different mode" is called a "debugger", and it's trivial to label stack entries or registers. That's what debuggers do.}

Doesn't the intermediate one disappear when the stack item is popped? Anyhow, the model in question is not intended to mirror typical production techniques, which are tuned for machine efficiency and not for human grokking. We've been over this already.


See also SubjectMatterExpert, DomainExpert, CommentOnCriticismOfTheGangOfFour, WhyFunctionalProgramming


OctoberThirteen


EditText of this page (last edited November 12, 2014) or FindPage with title or text search