3 ms·

Interesting, but correlation does not imply causation. Yet, unfortunately, that is the first response many will have to reading this.

• dllthomas 14y ago
  "Correlation does not imply causation" is not a meaningful reaction to an article like this. Yet, unfortunately, that is the first response many will have to reading this. This is more about a model than a study, and the model involves causation. Data fitting the model, therefore, implies causation! Not, of course, "implies" in the strict logical sense; it's possible that there is another model that lacks the causal link and also fits the data. But rather, "implies" in the sense that it's reasonably strong evidence - the more confident we are in the model, the more confident we are that there is causation involved.
• Symmetry 14y ago
  No, but correlation correlates with causation because causation causes correlation. Correlation between A and doesn't tell you if its (A->B) or (B->A) or (C->A and C->B)[1]. And robust correlation pretty much does imply that causation is happening somewhere. So then you have to look at the plausibility of each of those to try to figure out which it is. And since (assortative mating) -> (reinforced traits) has a clear and reasonable mechanism it has a serious leg up here. [1] Interestingly (A->C and B->C) doesn't cause correlation between A and B despite causing correlation between A and C and B and C. Which is how you can figure out causation from enough correlation datapoints.
  • dllthomas 14y ago
    > No, but correlation correlates with causation because causation causes correlation. Ow. Correct, but ow.