You seem to be talking about approaches in education. What about eg a government trying to decide on economic policy? Or relevant to the op, a car insurance company trying to decide whether to take sex into account when determining rates?
Are those decided by actuaries pulling ‘’‘theories’‘’ out of their asses or by actually looking at insurance claims?
I suspect they are based on data, rather than ass play. Are you saying nobody should be curious about why women drivers are less likely to be at risk for accidents/claims than men? I mean, we sort of know that men are more likely to be jackasses behind the wheel. Is it because driving school instructor teach men to be more aggressive while driving? Or is it something else? Should we find out or no? How deeply should we inquire?
False dichotomy. Theories are generally based on observations, and insurers’ risk assessments are based on observations of previous claims. Theories make predictions that can be disproved or not, and the insurer needs to correctly predict risk. They may not care about the ‘why’, but their algorithms do make testable predictions.
So, you’ve, yet again, presented us with something that has nothing to do with scientific theories.
I agree. “Scientific hypotheses”, would have been a much better subject title and topic of conversation, avoiding unnecessary distraction.
Perhaps the mods could change the title to avoid aggro?
You should PM them and ask.
Yeah, but, some of the things being brought up aren’t even “scientific”.
As said, not a scientific theory. Basing insurance rates, or any decision, on group attributes is not a scientific activity at all. Studying why sex/age/location/income/etc. would effect driving/health/etc. could be scientific.
How it is applied is a socio-political action that may or may not need to be regulated. If insurance companies could not take age and health into account, they would probably never offer life insurance that anyone could afford. Basing mortgage rates on race, on the other hand, would be a big problem because of the historic discrimination and the feedback effect involved.
Black people get higher interest rates and can only buy in areas with depressed property values therefore they have a higher default rate therefore they should pay higher rates and be barred from nicer neighborhoods. The policy creates and reinforces the conditions which justify the policy.
Science is a process. I know we hand it out like an honorific or a job title, but really, it’s a process. If it is followed down whatever alley researchers want to take, then it is “scientific”. Even if, maybe especially when, it proves the hypothesis to be incorrect. Scientists are wrong A LOT.
It would still depend on which category the relevant theory or hypothesis falls into. With regards to car insurance companies, the data does indicate that young men present a higher risk than young women. The hypothesis in this case would be that younger men are more likely to drive recklessly, and act more recklessly in general. In this particular case I have no problems with the insurance companies having a higher rate for young men. There is enough evidence to support the hypothesis, including a proposed underlying mechanism for why this is the case. Given that insurance is all about risk pools rather than individual actors, I have no problem in this particular case.* I go wherever the evidence leads, and if that means that young men are more likely to be reckless drivers because of testosterone, then that’s just the way it is. At this point I’d put the hypothesis that young men are inherently more likely to take risks than young women into the 1st category. I’m not going to deny that out of some misguided notion of trying to make everyone equal if they in fact aren’t. If there are legitimate differences, there is no point in denying them.
As far as economic policy goes, that would of course depend on which policies we’re talking about and what evidence is used to support those policies.
*. While insurance is all about risk pools rather than individual actors, most other things aren’t. Admissions to college, police stopping people, banks deciding whether or not to approve a loan, etc. All those things should be judged only on the individual in question, and group tendencies should not be taken into account, not even a little.
A distinction without a difference. When criminal vandalism and violence are used to thwart and sabotage experiments, theories cannot be adequately tested.
I can see how that could happen. Or the higher interest rates themselves could increase risk of default.
It might be more contentious if the risks had turned out to be the other way around, or to also vary by race. The EU has banned car insurers offering lower rates to women; they said it was sex discrimination)
A college application is far more individualised, but I’m not sure a loan is all that different to taking out insurance. Both the bank and the insurance company are concerned with the risk they may lose money. There may be an interest for society in barring the use of group memberships that people have no control over when calculating these risks, but that’s not the same as claiming they don’t exist.
Soviet Russia provides at least one example. Darwinian/Mendellian selection was considered too individualistic a process to drive change; Communist theory cried out for a collectivist explanation. Hence Lysenkoism and the imprisonment or execution of 3000 biologists who pointed out that Lysenkoism was horseshit. Hence also an enormous amount of wasted effort and a decline in agricultural output.
In my opinion the issue should be if there is an actual underlying biological difference. With regards to race, for example, cystic fibrosis and sickle cell anemia were mentioned earlier in the thread. These two conditions are linked to European and African ancestry respectively, and we have a good understanding of the genetics of both conditions. The evidence for that would fall into the well established and generally accepted by the experts as being correct. As far as I know there isn’t any serious scholars claiming, for example, that sickle cell disease is found equally in all racial groups.
Intelligence, on the other hand, doesn’t fit into the same category. There is no evidence to suggest an underlying biological mechanism for people of one racial group to have more or less intelligence than people of other groups. In fact the evidence that we have suggests that the differences we observe are due to differences in opportunity and culture rather than inherent biological differences.
That’s pretty well said.
Which leads to two thoughts.
Thought 1:
Group membership is used as an inexpensive (=lazy) proxy for individual specifics. If I have the applicant’s credit report I don’t need to look to their race or sex to make an informed credit-granting decision, and in fact race and sex would be, statistically speaking, confounding variables given the more specific data in the report.
But if I don’t have or can’t get the report, using race or sex is statistically better = more accurate than using a coin flip. For the bank. But will be perniciously worse for some subset of the individuals so treated. So whose “rights” should prevail: lender or borrower? That is a political question, not a scientific one. Don’t try to make it into a scientific one; that’s a category error.
Thought 2:
As to people and groups, the width of the distribution and the overlap of the distribution are highly significant parameters. Choosing to use a statistical property for an individual decision can potentially be statistically defended if the distribution is tight enough and the overlap is small enough. But cannot if the opposite is true.
Said another way, in the former case the group property has high predictive power. In the latter it has small to nil predictive power.
Which leads us back to the political question. If decision makers use group memberships with small predictive power, they will make statistically better = more accurate decisions than using a coin flip. But only slightly better overall than coin-flips. But those same decisions will be perniciously worse for some much larger, potentially huge subset of the individuals so treated. IOW, the smaller / fuzzier the effect the greater harm its use will cause vs the smaller benefit given.
Which suggests a rubric for when using group attributes might be deemed “scientific” enough to be politically defensible in a democracy. The group attribute must:
- Be precisely and accurately measurable.
- Have a tight distribution of members.
- Have wide separation between the bucket values of that attribute.
- Have a very high statistical correlation to the managed outcome.
- Not be used in the presence of other more specific or more individual data available.
IMO the OP desperately wants to believe there are some human group attributes that easily meet these hurdles that mainstream society is greatly afraid of. Such that society has its collective eyes screwed shut to the "facts" obvious to the OP. To such an extent that society is actively suppressing knowledge of and research about said "facts".
The OP thus far has lacked the courage of their convictions to do more than JAQ and hint around the edges of those beliefs.
A good debate is about a concrete topic. Perhaps we should finally be told what that concrete topic is.
I cannot put it better than Bertrand Russel:
"When you are studying any matter or considering any philosophy ask yourself only what are the facts and what is the truth that the facts bear out. Never let yourself be diverted either by what you wish to believe or by what you think would have beneficent social effects if it were believed, but look only and surely at what are the facts. "
No, they’re not.
They’re more prevalent in Northern European populations, and a belt of people that includes parts of Africa, the ME and India, respectively. That those happen to be the groups that are the biggest contributors to European- and African-American gene pools doesn’t make the quoted statement accurate at all.
As I have to point out so frequently in these threads: Zulus and Bushmen don’t get sickle cell.
I don’t think the example is a good fit with your OP, because AFAICT none of the “Hindutva” opponents to the hypothesis of ancient Indo-European in-migration to South Asia are advocating suppression of any scientific theories involved with ancestral-DNA analysis.
Some of them may be opposing or misinterpreting the findings of this particular instance of research, but AFAICT nobody is saying that such research shouldn’t be done, or that the theory should be suppressed, on the grounds that it’s too harmful despite being admittedly true.