I disagree. Take the education reform problem (in the US). A number of solutions have been proposed, which one should we implement, or should we come up with a new one? Are teachers' ratings completely bogus, as some claim, or can they be used to rate them. People want answers, you can't just give up. Not everybody can close shop and work on other, more well-defined problems because these problems need to be solved, or lacking that, addressed. You have to come up with a methodology. Now, you may not want to call it "science" but that doesn't make the results meaningless.
Similar problems that require a combination of statistics, biology, psychology, ethics, and politics:
* Should universities stop affirmative action, should they stop reverse affirmative action for Asians?
* Should we (like Norway) enact laws that require a certain percentage of women on company boards?
* Should there be gender equality (in the sense of numbers) in almost all professions? Are men (or women) more suitable for certain jobs?
I think you're completely wrong about education. The science is already there. We have empirical evidence on ways to improve it. The problems with education are almost all political... but this thread isn't really about education, though is it? From reading your posts, it seems like you really want this thread to be a battle over education and if I had to guess I'd say that Feynman's insightful almost throwaway comment on education really struck a nerve.
> Now, you may not want to call it "science" but that doesn't make the results meaningless.
If you're not doing good science, then your results will likely only be indirectly useful. The last 50 years of increasing education expenditures without showing any results speaks volumes about what a tragedy having an attitude like yours is. Maybe if someone had listened to Feynman in 1974 and quickly eliminated any educational expenditures that didn't produce results, we wouldn't be having this argument.
You can do meaningful science with people. There are plenty of solid drug studies out there, and even good education research. What I mean is there are experiments you don't get to do with people and when that happens you need to move on and not try and get by half assing it.
So, you don't get to see what happens when people eat the exact diet you want them to for 20 years. Having people fill out servery on what they ate is not a reasonable substitute. In this case there are options like sending people actual food. You don't know if they eat it, but you do know you sent it and you can measure outcomes based on what food was sent.
PS: Also, most of what your talking about is almost pure politics. When politics and science really clash it's rare for science to make much headway. See: Climate Change, Evolution, etc.
Similar problems that require a combination of statistics, biology, psychology, ethics, and politics:
* Should universities stop affirmative action, should they stop reverse affirmative action for Asians?
* Should we (like Norway) enact laws that require a certain percentage of women on company boards?
* Should there be gender equality (in the sense of numbers) in almost all professions? Are men (or women) more suitable for certain jobs?
etc., I'm sure you can think of many more.