[rsanchez1]

The author didn't pick good examples of contrarian anecdotes found on HN. They are contrary to a study on coffee consumption and dementia. The study, however, is only a sample of 124 people from Tampa and Miami. The study is just as likely to have found an isolated effect as the posters with family members who consumed a lot of coffee and still developed alzheimers.

You would be better served by this:

https://sites.google.com/site/mccormickphilosophy/home/criti...

The author must be a frequent coffee drinker who didn't like that some people had a different experience with coffee than some other people had, and felt compelled to write that post. It's not the contrarian anecdotes that left me with the sense that the research findings weren't conclusive. It's the fact that a population of 124 people in two cities is not at all representative of the target population which numbers over 40 million according to the census bureau.

http://quickfacts.census.gov/qfd/states/00000.html

Maybe when the study researchers conduct a larger study I'll believe them.

[tylerhobbs]

Great comment. I am most certainly not claiming that you shouldn't question research findings on other grounds, such as sampling bias or statistical significance, as you have done here; those types of points are extremely valuable.

[scalable]

N=124 can definitely give statistically significant results. Consider the following thought experiment. You have a coin that you think might be loaded. You flip it 124 times. You get 119 tails. It is probably loaded. Lets say you get 66 tails. Could definitely be due to chance. Moral: The larger effect you study, the less sample size you can get away with. It might be that people from Tampa and Miami are different in some important way from other people. However, if you want to level this criticism, maybe you should give a reason for suspecting this? "The study is just as likely to have found an isolated effect as the posters with family members who consumed a lot of coffee and still developed alzheimers." No, no, triple no.