Hacker News Reader

Comment by powera | Hacker News Reader

My personal take: a demonstration that the burden of proof in a scientific (or technological) sense is very different from the burden of proof in a legal sense.

View on HN

This has always been the case. The standard for scientific proof in the natural sciences is often quite high, and the legal standard for tort claims is usually "preponderance of evidence", which you could think of as a P-value of 0.50. The lowest P-value acceptable in most scientific fields is 0.10, more common you would use 0.05, and particle physicists would use 0.0000003 (for announcing the discovery of a new particle).

p-values are B.S. unless used by a statistician.

This statement is obvious nonsense. You might as well say that chemical names are B.S. unless used by a chemist.

Used by a chemist, BS would be B2S3. You have to balance the charges to use BS properly.

p-values are especially dangerous if used by a statistician. They'll know best how to manipulate the charts and data to look good and fool laypeople. Motive and intent are the most important issues.

"Scientific conclusions and business or policy decisions should not be based only on whether a p-value passes a specific threshold.

P-values can indicate how incompatible the data are with a specified statistical model.

P-values do not measure the prob. that the studied hypothesis is true, or the prob. that the data were produced by random chance alone."

Source: http://amstat.tandfonline.com/doi/pdf/10.1080/00031305.2016....

Maybe the US should do what the EU does and ban such substances the moment there is some proof that they may do harm, and the burden should be on the industry to prove without a doubt that they don't cause harm.

Instead, the US government is trying to do the opposite by repealing such policies in the EU and making them more like in the US where anything is accepted until millions get cancer from it.

Water may do harm (it causes drowning). Oxygen may do harm (it increases the intensity of fire).

"May do harm" cannot possibly be the standard.

> the burden should be on the industry to prove without a doubt that they don't cause harm

This is incoherent. There are no substances like this.

Exactly. This doesn't make any sense. Sunlight causes cancer. Maybe we should start there?

> Maybe the US should do what the EU does and ban such substances the moment there is some proof that they may do harm,

We know that processed red meat causes cancer. Processed red meat has not been banned by the EU. There are countless other examples.

That is just not true. The red meat part (http://chriskresser.com/red-meat-cancer-again-will-it-ever-s...). Also talc is legal in EU.

What? It is true. We know that processed red meat definitely causes cancer. It doesn't cause very much cancer - the risk is very small.

http://www.europarl.europa.eu/RegData/etudes/ATAG/2015/57130...

Your link starts with hyperbole (no one is being told not to eat meat)

Why are you telling me that talc is legal in eu? That just agrees with my point that eu does not ban things that we know are dangerous.

> It doesn't cause very much cancer

I can tell you're an expert on the subject from the very technical language you use.

EDIT: holy fuck IHBT.

If you think writing like an orc is "audience appropriate", you're mistaken.

Here's a preliminary release on the original study: http://www.thelancet.com/pdfs/journals/lanonc/PIIS1470-2045(...

They picked a 95% confidence level, which means they accept that 1/20 of their positives will be false positives.

They did this when conducting a study on "more than 15 types of cancer".

These people are about as smart as you are.

Reply on news.ycombinator.com