Hacker News Reader

Comment by therein | Hacker News Reader

It might be helpful to mandate every submission to an academic journal to be accompanied with a maximum 800-word "layman's summary" that will be directed at the media and the public. Something more comprehensive than the Abstract/Preface without any terms that wouldn't be known to the general public.

View on HN

A lot of journals are starting to do this. For example PNAS (a top ranking journal) introduced a required "significance statement" in addition to an abstract, which is meant to be understood by a more general audience. So do the PloS journals, and many others.

http://www.pnas.org/site/authors/procedures.xhtml#significan...

http://journals.plos.org/ploscompbiol/s/submission-guideline...

PLoS says "The goal is to make your findings accessible to a wide audience that includes both scientists and non-scientists", and PNAS says it is meant to "explain the relevance of the work in broad context to a broad readership."

This sounds like a good idea. But, it may not be possible in most of the hard sciences. Imagine writing an 800-word layman's summary for some results in string theory.

By 'most of the hard sciences', you seem to have chosen a sliver of one of the hard sciences, which plenty of people in that science think is just smoke and mirrors. If a 'general audience abstract' is a good idea, it shouldn't be balked at because of a single theoretical field which is quite controversial. Let other journals handle those papers :)

Unless string theory all of a sudden gained everyday applications i think i would aim for fields like biology, chemistry and similar first.

Yes, biology makes the most sense, as thats usually what gets "translated" to clickbait style headlines like "Quinoa cures cancer".

This week i read a headline that could be interpreted as saying green tea would prevent zika infections.

In reality it was a chemical of some sort found in green tea (and elsewhere) that seemed to counteract zika. But it required concentrations that you would never find in nature.

I'm personally against such an idea:

- The journalist isn't supposed to be a copy-paste robot. Even if the press release contains big claims, the journalists should be blamed if they don't do fact-checking. An example from the recent times (and not scientific): If the conman is claiming to be a bitcoin founder, the serious journalist aren't supposed to just repeat his claims.

- The "layman summary" as in "relevance" is impossible to be realistic for most of the basic research. Is research relevant if it can be applied in 10 years? In 20 years? 100 years? How can you even estimate when it can be applied? The realistic estimate of the application of the research is possible only if the research is targeted at the short-term goals. Which is then more engineering than science. But if you include long term view, then what could be "possible" is typically not what the "public" understands ("but you said we will be able to use x and where is it now you liars.")

- The scientists should actually be somehow also motivated not to exaggerate the applicability of their research, as a force against their reasonable and real motivation to do so. We should be able to let the researchers do the research even if we don't expect the immediate "use." Otherwise we don't support the science, just solving the shortest-term problems.

I can give you the layman summary for almost every study that involves any kind of statistics:

"Is this a single study or a systematic review / metaanalysis? If it's the first then ignore it. It's more likely false than true."

There was some recent analysis indicating that, for psychology, the correlation between the first study of an effect and the eventful meta-analysis result was around zero, i.e.,such that an initial result conveys zero new information about an effect. That is fucking mind boggling.

Reply on news.ycombinator.com