Hacker News Reader

Comment by epistasis | Hacker News Reader

I'm not arguing that the IARC made the right scientific call on this monograph, and in fact I think they are likely to be wrong since so many other organizations disagree. However, I think it's absolutely essential to protect the private deliberations in science.

What should be open is the final reasoning, the results and data that go into a conclusion. But the process of getting there should allow people to make all sorts of mistakes until they get it right, without facing undue scrutiny during that process.

There are far too many biased groups that will take any indication of a simple arithmetic error that gets corrected as a sign of corruption or manipulation.

In mathematics, a proof shows all the work, but you don't have to show all your failed attempts until you get to a proof. I don't think that these scientists should be treated differently.

View on HN

Generally, I think this is true. If a manuscript goes to review before publishing, typically the reviewers are anonymous and the reviews are shared only with the authors of the submitted manuscript. This process is usually secret and does protect the ability to be candid in review. The final published manuscript stands on its own atop its data, cited sources, and arguments.

We could have a debate about whether this is a good way to do things, and surely it could be improved, but it has worked not too badly in the past.

I think this is a little different though. The IARC is publishing a review of the literature, but it is both the author and the publisher. As such there's no intermediate, outside and independent step, and they should be extremely transparent about how they came to their conclusions. They should have reams of studies cited, and they should detail exactly how they weigh contrary strands of evidence. Simply saying, "it was sponsored by industry" is an inappropriate reason to eliminate a study. Again, each study stands on its own data, cited sources, and arguments. Be more critical if you don't trust the source, but consider the data and arguments.

I'm struck by how the IPCC seems to operate by casting a wide net for experts and striving to build a consensus everyone can get behind, and sign their name to. Shouldn't the IARC work the same way?

If the reasoning process is not thoroughly detailed in the original monograph, they absolutely should release an anonymized version of the deliberations.

The thing to remember as well with IARC is that it does its classifications with a hazard based approach, while most regulatory agencies use a risk based approach.

A hazard is something that can harm potentially harm you. Risk quantifies how likely that will happen.

Edit: wording

Making the "final reasoning" open without also making the data that was used to reach that final reasoning creates too much room for speculation and conspiracy theories.

Show me the data!

If I can examine it and reach the same or a different conclusion, then we can begin a real discussion on why. That's science.

The data are public. http://monographs.iarc.fr/ENG/Monographs/vol112/index.php

I don't think anyone was arguing the data that the final evaluation is based on should not be open.

The point is that the intermediate drafts of the evaluation itself do not need to be public and it's fair to deliberately keep them in the circle of people who are working on it or reviewing it.

There is no such thing as "undue scrutiny" for a public policy group. There is no amount of scrutiny that is too high. If you don't want scrutiny, don't take public money or make public policy.

Reply on news.ycombinator.com