I’m choosing to ignore that possibility to clarify the exposition of what I think is going on. Problems like that are what I’m referring to when I preface:
And we can’t explain all of this away as the result of illusory correlations being throw up by the standard statistical problems with findings such as small n/sampling error, selection bias, publication bias, etc.
Even if we had enormous clean datasets showing correlations to whatever level of statistical-significance you please, you still can’t spin the straw of correlation into the gold of causation, and the question remains why.
You could say that “A and B happen to be in sync for a while” is possibility 3, where C is the passage of time. (Unless by “happen to be in sync for a while” you mean that they appear to be correlated because of a fluke.)
To generalize, it’s also possible that you’re observing survivor effects, i.e., both A and not B (or B and not A) cause the data to appear in your data set.
There are at least two more possibilities: A and B are unrelated, but happen to be in sync for a while, and the data was collected wrong in some way.
I’m choosing to ignore that possibility to clarify the exposition of what I think is going on. Problems like that are what I’m referring to when I preface:
Even if we had enormous clean datasets showing correlations to whatever level of statistical-significance you please, you still can’t spin the straw of correlation into the gold of causation, and the question remains why.
You could say that “A and B happen to be in sync for a while” is possibility 3, where C is the passage of time. (Unless by “happen to be in sync for a while” you mean that they appear to be correlated because of a fluke.)
To generalize, it’s also possible that you’re observing survivor effects, i.e., both A and not B (or B and not A) cause the data to appear in your data set.