Hacker News Reader

Comment by dalbasal | Hacker News Reader

Idk the exact answer, but I doubt the reason is "researchers don't understand the statistics." It is some "how the sausage is made" reason, most likely.

For example: researchers have the budget/resources to to a preliminary study. They use the interesting results of this study to secure funding for a larger sample size study.

Btw, a small sample size study isn't useless necessarily. You can refine methodology, with less on the line. You can also (more or less) validly exclude certain conclusions, to avoid wasting resources on a wider study.

The problem is with pop science headlines, like it is in any other journalistic field. I think it's less of a problem though, than many think. Publishing in the popular press needs to be understood as "academics working on X" not "X has been proven."

View on HN

>The problem is with pop science headlines, like it is in any other journalistic field. I think it's less of a problem though, than many think. Publishing in the popular press needs to be understood as "academics working on X" not "X has been proven."

Writing a headline that misrepresents a single study's conclusion as "proven" or what "scientists say" -- especially when the conclusion is contradicted by most others on the subject -- should be considered by journalistic ethics watchdogs as an extreme violation bordering on an outright lie.

The media seems simultaneously bewildered as to why the public no longer trusts it, while incapable of recognizing and responding properly to obvious breaches of the public trust like inflated science headlines.

Its not just science, it's anything. Also, it's not just media, it's everything.

The only thing to do is individual. Take "facts" with a grain of salt, at the very least until you have read the article. Do not take headlines as facts.

You can just read academic journal abstracts, if you want a more academic and precise reading materials.

There are trade offs between rigour and the digestible content intended to be browsed over coffee.

>Take "facts" with a grain of salt, at the very least until you have read the article. Do not take headlines as facts.

One consequence of social media is that headlines now have outsized importance in relation to the substance of actual articles.

Headlines, thanks to links and embedded snippets, get dispersed to 1000x more eyeballs than the articles themselves. So when they contradict or misleadingly oversimplify (which is often, especially since they are often written to draw clicks) it's the headline version of the truth that wins out, while the actual readers are left to chime in with "Hey, that's not what the article actually said..." in the comments below, swept away in a sea of respondents who commented only after reading the headline.

Journalistic ethics haven't caught up to this. They're still playing by the rules of legacy media, where you can bend the truth a little to make a joke or to draw people in. The problem is that in the digital age, that unfortunately amounts to spreading a falsehood that usually overshadows the facts of the piece.

Let’s go one step further - unless you have some experience understanding academic journals (many here do), even the journals may suggest something that isn’t true by way of you reading into it.

Some things everyone here should know (like correlations and what they mean). Other things many won’t know, like certain statistics methods, base rates for a given thing, etc.

It’s easy for two people to read the same study and come out with totally different conclusions based on their individual context.

Reply on news.ycombinator.com