4 ms·

>The next time someone asks us to trust the science, our response should be "okay, but can I trust the scientists?" Originally, trust in scientists was not req

by forcry 5y ago

>The next time someone asks us to trust the science, our response should be "okay, but can I trust the scientists?" Originally, trust in scientists was not required. Because anyone could have repeated the experiment and should see results consistent with what the scientists were saying. This is where the credibility of science come from and it made people's trust in science grow to a kind of blind faith. But then some point, experiments became too large to be attempted by anyone. This along with entities like "consensus" and "reputation", that can reject any conflicting results, enabled entities that control all of these to have a veto power on scientific truth. This is our current situation. And I am not sure people recognise the true danger of this situation. This is the kind of power that in holy men held in ancient times, where you could make people to do mass murders and other atrocities in the name of god. So I think it is very dangerous. It might not be that bad as of now, but we are in that path. One solution to this problem is to recognise and reduce the trust in "science" that is a result such collaboration. May be we can do it by using a measure of how easily verifiable a result is. For example, if something is verifiable by anyone in this planet, it should have a measure of 1. And if it is only verifiable by a single entity, then it should have a measure of zero. If it is verifiable by a few entities, then it should be somewhere between.

native_samples 5y ago
That's a nice idea but there are a lot of problems with it. 1. Being able to replicate a study is necessary but not sufficient for the conclusions to be correct. Scientists pump out a lot of studies that are just doing stats on public data sets. These should be easily replicable, but the conclusions may still be nonsensical if the analysis isn't done right (and a lot aren't done right), or if the scientific work itself contains fraud or logic errors (and a lot do). 2. There are no incentives for scientists to make their work replicable. Their salary does not depend on it. Their salary isn't going to depend on it, because the scientific institutions are all basically dependent on funding by governments or rich benefactors who aren't paying attention - funding "science" is not a means to an end but the end in and of itself for these groups, and thus they aren't particularly motivated to double check the quality of what gets released. Consider the above problem: scientists were deceiving people because they thought helping Trump is the same thing as helping "racists". Scott Alexander did a very thorough analysis of this where he pointed out that Trump is not, in fact, racist and a lot of the claims that his followers are don't make any sense [1]. If self-proclaimed scientists can't do even basic sanity checks on their own beliefs like that, you can imagine how much they struggle with more complicated claims. [1] http://slatestarcodex.com/2016/11/16/you-are-still-crying-wolf/ http://slatestarcodex.com/2016/11/16/you-are-still-crying-wo...
- forcry 5y ago
  >Scientists pump out a lot of studies that are just doing stats on public data sets. May be I didn't make myself clear. I only meant to include actual experiments, in the context. So it will only involve procedures that include collecting raw data from nature. Mostly agree with the rest of your comment.
  - native_samples 5y ago
    Yes, this gets into the (very important!) topic of what exactly the definition of science is. Consider that the normal 'street' definition of people doing experiments on nature would mean the following fields aren't scientific: 1. Computer science 2. Economics 3. Climatology I'd personally be fine with all three being considered un-scientific or even defunded, even though I quite enjoy reading computer science papers. It's more of an engineering discipline in my view and the attempts to make it seem scientific mostly just obfuscate things. And most people would be fine with economics not being a science. The live wire is of course climatology. These people don't do experiments, they just compile data and analyze it. Good luck convincing people that climatology isn't a science, even though you may be in the right about it!

by forcry 5y ago

native_samples 5y ago

That's a nice idea but there are a lot of problems with it. 1. Being able to replicate a study is necessary but not sufficient for the conclusions to be correct. Scientists pump out a lot of studies that are just doing stats on public data sets. These should be easily replicable, but the conclusions may still be nonsensical if the analysis isn't done right (and a lot aren't done right), or if the scientific work itself contains fraud or logic errors (and a lot do). 2. There are no incentives for scientists to make their work replicable. Their salary does not depend on it. Their salary isn't going to depend on it, because the scientific institutions are all basically dependent on funding by governments or rich benefactors who aren't paying attention - funding "science" is not a means to an end but the end in and of itself for these groups, and thus they aren't particularly motivated to double check the quality of what gets released. Consider the above problem: scientists were deceiving people because they thought helping Trump is the same thing as helping "racists". Scott Alexander did a very thorough analysis of this where he pointed out that Trump is not, in fact, racist and a lot of the claims that his followers are don't make any sense [1]. If self-proclaimed scientists can't do even basic sanity checks on their own beliefs like that, you can imagine how much they struggle with more complicated claims. [1] http://slatestarcodex.com/2016/11/16/you-are-still-crying-wolf/ http://slatestarcodex.com/2016/11/16/you-are-still-crying-wo...

forcry 5y ago
>Scientists pump out a lot of studies that are just doing stats on public data sets. May be I didn't make myself clear. I only meant to include actual experiments, in the context. So it will only involve procedures that include collecting raw data from nature. Mostly agree with the rest of your comment.
- native_samples 5y ago
  Yes, this gets into the (very important!) topic of what exactly the definition of science is. Consider that the normal 'street' definition of people doing experiments on nature would mean the following fields aren't scientific: 1. Computer science 2. Economics 3. Climatology I'd personally be fine with all three being considered un-scientific or even defunded, even though I quite enjoy reading computer science papers. It's more of an engineering discipline in my view and the attempts to make it seem scientific mostly just obfuscate things. And most people would be fine with economics not being a science. The live wire is of course climatology. These people don't do experiments, they just compile data and analyze it. Good luck convincing people that climatology isn't a science, even though you may be in the right about it!