[rawgabbit]

Minor correlation? P values are small indicating a strong correlation?

      Quote:  Of the bacteria detected, oral viridans group streptococcal DNA was the most common, being found in 42.1% of coronary plaques and 42.9% of endarterectomies.  Immunopositivity for viridans streptococci correlated with severe atherosclerosis (P<0.0001) in both series and death from coronary heart disease (P=0.021) or myocardial infarction (P=0.042).

[teraflop]

This is a super common misconception, but a small p-value does *not* (necessarily) mean a strong correlation. It means high confidence that the correlation is non-zero.

[rawgabbit]

Yes. I understand that. I was questioning the original assertion that there was a “minor” correlation. The p values indicate a correlation. One that is statistically significant.

[teraflop]

Late response, but the whole point is that "statistically significant" doesn't necessarily mean "major", because significance is as much about the experimental design as it is about the thing you're measuring.

Imagine, for the sake of argument, that being left-handed is correlated with a 0.001% higher likelihood of accidentally dropping your car keys every time you pick them up. An experiment studying a small sample of people for a month probably wouldn't detect this correlation at all. If you ran the same experiment but carefully monitored every single person in the US around the clock, you would be able to reliably detect it with an extremely small p-value. And yet it's still fair to describe it as an extremely minor correlation.

[raincole]

P value means "if the null hypothesis were true, the probability we would have observed what we actually observed."

It's definitely not the strength of correlation. It's not even the probability that the opposite of the null hypothesis is false!

[cluckindan]

That’s what it means in the literal sense. As a more practical interpretation, p value is the ”probability that the observed result was due to random chance instead of the suggested hypothesis”

[drgo]

.. and only if myriad of other assumptions (many are unstated) hold.