Is there any correlation or causation here? I have the following data, where 2 properties (P1 and P2) can be either True or False
              P1      P1 
             False  True

P2:False    2646    400
P2:True     749     245


I am trying to understand if P1 being True or False has any implication on P2. At first sight, it seems that if P1 is True, there are better chances that P2 is True. But I am sure that this is too naive.
 A: *

*Is there any correlation? Using  2x2 contingency table analysis, the Phi coefficient of association is 0.14, with chi-squared = 74.08 and n = 4040.  So, yes, there is some correlation (or association). 

*Is there any causation? I don't think it will be easy to determine that philosophically, statistically, or otherwise. There is simply not enough information to know how these numbers came to be. I think you need to see if this data is observational or experimental; some people seem to believe and/or argue that you cannot make causal claims from using observational data. I think you need to pick a causal framework (one that defines what is causality) and see if the causal calculus (if any) available in that framework will allow you to draw a causal link between these two variables. 


Note these sayings


*

*Correlation does not imply causation.

*Causation leads to correlation. 


So, it seems to me, a causal relationship should manifest (or cause) a correlation relationship. But, just because you see a correlation relationship, does not mean there is a causal relationship. Kind of like (as an intuitive example), fire leads to smoke, but observing smoke may not mean there is fire. 
So, I believe, you may have a causal relationship, it is not ruled out yet, because you have a correlation relationship. But it is not definitive, whether you have or do not have a causal relationship. Remember, stock prices have been shown to be correlated with moon cycles, yet, is there a causal relationship? 
