Hacker News Reader

Comment by tokenadult | Hacker News Reader

We have simply got to stop taking submissions of preliminary research findings from low-impact factor primary research journals for discussion on Hacker News. Let's first of all see a neutral review of the issue by an active scientist or (better still) a competent science journalist (that's a rare kind of person) to see what of methodological problems there may be in this study, and how well it represents the prior literature on the topic.

View on HN

I'm not really qualified to comment and I didn't spend very much time analysing the paper. However, I'll comment a bit since reading nutrition papers used to be a hobby of mine. First, the sample size is relatively good - nearly 1000 people, taken from a similar area and over a long period of time (30 years). What's missing is any obvious mention of how the participants were chosen, so it's hard to see if there is any selection bias. However, I didn't look at the links provided, so hopefully it goes into more detail there. I mention this because I have often seen papers published in this area with obviously flawed sample selection. This seems better than most.

The main things that stand out for me is that they did not attempt to measure how much chocolate people consumed, nor the type of chocolate consumed. They simply measured the number of times a week "chocolate" was consumed (which could have been a large variety of things). The other main point is that they mention that there was a significant reduction in alcohol consumption in the groups that ate the most chocolate, however they did not try to correct for this. Just scanning through the data they present, I don't think it is possible to tell if the increased cognitive performance is related to chocolate intake, or a lack of alcohol intake.

So, it is an interesting paper, but I wouldn't go out and start eating chocolate because of it.

I disagree. First of all "Appetite" for which this article was accepted is decently ranked (whatever you think of these rankings). It's in Q1 for both Nutrition and Psychology. H-Index of 80 is pretty solid. It's the 179th ranked Psychology journal which isn't bad for such a niche journal. For Nutrition it's ranked 23rd [1]. So as far as food related cognitive science is concerned I'd call this a pretty decent journal. Since it is accepted it has gone through a successfull review process already.

More importantly though...why is it bad that unpublished research is linked to on HN. I'd rather read it and make up my own mind than wait for it to go through a review process. The topic at hand is at least tangentially related to core HN content, too as eating chocolate might make you a better programmer/entrepreneur.

[1]: http://www.scimagojr.com/journalsearch.php?q=29994&tip=sid&c... It has a (whatever you think of these is another matter)

Reply on news.ycombinator.com