But in practice, science depends on the trust that scientists aren't actively subverting the process: failing to consider alternate hypothesis, fabricating data, ignoring data, deliberately filtering for confirmatory data, groupthink, etc.
So an import safeguard is the ability to detect when someone is doing the above, even if the detector doesn't know the literature well enough to refute it.
With that said, that doesn't mean there's enough evidence of misconduct to justify a subpoena in this case; I'm just saying, it's not simply a matter of "just look at the data rather than the people".
Using a subpoena to threaten scientists for finding results you don't agree with is a form of censorship.
So now what? How do you collect and analyze data when trust in those doing the collection and analysis is not a given?
One approach is to implicitly trust scientists, because they are experts in their field, while you are a novice. This is what a lot of us do, and I think it's a perfectly fine solution, but it is a trade-off; you might misplace your trust.
Another approach is to train up on the science yourself, so that you can independently verify things. This is a great solution, but it isn't scalable. Most people have other jobs and other interests, and in any case, you can never be an expert in everything that matters.
Another approach might be to try to build trust using out-of-band signals, like transparent communication. Transparency often sounds nice, but it also has a major chilling effect, because people communicate more efficiently in the sorts of candid conversations that are hard to have publicly.
Personally, I think the witch hunt for biased climate scientists is disgusting and beside the point, but I also find it distasteful to suggest that it isn't a really hard problem to disseminate trusted information about complex subjects in a highly specialized society.
> The Office of Technology Assessment (OTA) was an office of the United States Congress from 1972 to 1995. OTA's purpose was to provide Congressional members and committees with objective and authoritative analysis of the complex scientific and technical issues of the late 20th century, i.e. technology assessment. It was a leader in practicing and encouraging delivery of public services in innovative and inexpensive ways, including early involvement in the distribution of government documents through electronic publishing. Its model was widely copied around the world.
It was de-funded "Contract with America" period of the 1990s.
Science is global endeavour based on data. If one scientists is lax or fraudulent, there are numerous incentives for competing scientists to prove so.
I certainly agree that things should be open sourced, probably allowing for some kind of embargo system for publication, but we shouldn't allow a rhetoric to develop suggesting that people should be making things easily replicable, because that essentially means making very portable and resource-flexible code, which is not trivial, even in less complex domains, as I'm sure many people on this forum recognise.
As an example, many models make use hundreds of cores via MPI (along with preposterous amounts of RAM), and whilst I am no expert, I get the impression that porting such things to more commodity hardware is essentially impossible.
As an aside, I have seen more bad-faith practice (i.e. manipulation of results) in industry than in science, and have spent approximately equal periods of time in both.
http://www.gfdl.noaa.gov/cm2-5-and-flor
However, as mrow84 points out, you will need a supercomputer, unless you have a lot of time on your hands.
Are you serious ? Not only is there multiple way to interpret same data (after all there are so many competing models to explain it) and at best you can debate confidence intervals and models - it's far beyond non-experts expecting them to review the data directly is ridiculous - the only way they can work with experts is based on trust and credibility.
And no, it's not unreasonable to expect a non-technical person to review the data or the methods either. Many people don't want to put the time in to try but it's well within the reach of any high-school grad.
Not if you don't trust the field. It's like saying if you don't like your numerology readings hire a different numerologist and compare the results.
Not saying I agree with the view but science is a lot less convincing when when you can't setup repeatable experiments.
>And no, it's not unreasonable to expect a non-technical person to review the data or the methods either.
You're going to review PHD level work that ran weather simulations on supercomputers and compiled various data sources with highschool level knowledge in your spare time ?
Perhaps you could do a security audit of OpenSSL if you have some free time on your hands - I'm sure people will even pay you if you manage to find security issues and you' be doing the community a big favor. Maybe we should hire some highschool grads to do it - start a kickstarter campaign and I'll pitch in.
To be fair, this is excellent advice, much better than "your numerologist knows what he's talking about, and you'd better follow his recommendations".
No real reason to compare the results afterwards, though.
But sure, if you don't trust someone from the 'warming' camp, get a 'status-quo' camp person to check their work.
And as for numerologists - try looking for another person skilled in your actual problem, relationships, work, etc...
When a mathematician checks their work they do so by a different method, not just repeating the same steps to reach the same potentially flawed answer.
If you think that stuff like economics/climatology/etc. has the same level of rigor and validation applied to it as established and verified theories from hard science then boy do I have some bad news for you (struggling for SD 2 as ideal vs SD 5 for eg.).
Reading other people's mails won't refute the physics.
And no, other factors aren't the answer:
http://www.bloomberg.com/graphics/2015-whats-warming-the-wor...
It's the same with the Earth. It's as isolated as the isolated room would be -- we know exactly how much heat can dissipate and how much must contribute to the increase of the temperature for the room, and we have quite good estimates for Earth too (less heat dissipates to the space when CO2 increases, and the Sun shines the heat nevertheless - we can nicely calculate even the contributions of the changes in orbit and some clever scientist did that even before the computers were available, it took years of calculations by hand then but it was done). We can be wrong for some particular year, but the direction and the average increase match.
The room is of course simpler than the Earth. Imagine having some water pumps, some fans and some chunks of ice in the room. You know the formulas for the water and ice but you have more to calculate for short term effects. But you still know there will be less ice as the time goes by and more water. And still always a bit warmer.
CO2 is a green house gas - good job mr. high-school physicist - now show me the temperature projections along with human CO2 production graphs and estimated impacts and error margins and past model prediction performances vs measured data. You know the actual things that matter for policy implications. See that's where things go beyond high school.
http://www.ipcc.ch/report/ar5/index.shtml
There are other countries among those who took part in making the reports who would be more than eager to point on anything that's wrong there. And nobody did find anything significant against.
Wait - but you didn't verify it your self ? After all :
>You don't have to be a rocket scientist to repeat most of the basic calculations which give similar results.
Or does it all boil down to trust and your original argument is nonsense.
Edit: For those that are suggesting that you need the emails to verify that scientists aren't making up the data, the data itself includes information on its providence. All the information you need to trust the data come from independent verification of the data and from information in the data itself. You don't need emails to verify research, but if you did then there's a problem with how research is published.