Learning how to build a nest plays an important part in the breeding success of birds. For example, Dr. Snow has reco...

hales July 26, 2019

And to take the question one step further- this question is a cause and effect argument: I've read through the other posts but still am murky on what the relationship is between the correlation/causation here and the cause/effect argument that the author is making.

I think where I am struggling is I thought you are not supposed to assume causation from correlation.

I've watched this part in the lecture several times too, but just can't seem to wrap my head around it. How do you work from one to the other?

Irina July 26, 2019

@hales,

Let's look at the question first.

The author argues that because breeding success appears to correlate with the nesting experience, nesting experience must improve blackbirds' breeding success.

How can we weaken this argument?

I am sure you are familiar with the saying that "correlation does not imply causation," which in logic terms means that correlation is not a sufficient condition for causation, but in statistics, correlation actually often gives us a hint that there is a casual relationship between two variables. When I studied this topic in statistics, one of the lectures summarized it in this quote:

"Empirically observed covariation is a necessary but not sufficient condition for causality."

Correlation thus might suggest that there is causation or not, we cannot conclude either way as you correctly pointed out, but when there is causation, there is correlation as well. When we see correlation between two or more variables, how do we know what is the true relationship between these variables?

Let's use the example from the passage -breeding success appears to be positively correlated with nesting experience. The true relationship might be any of the following:

(1) A causes B
Nesting experience causes breeding success.

(2) B causes A
Breeding success causes blackbirds to build better nests (perhaps, they learn from their mates?)

(3) A & B are consequences of a third confounding variable, but do not cause each other;
Blackbirds that live in a certain area tend to build better nests and have greater breeding success, thus geographic location is a third variable.

(4) A causes B and B causes A
Nesting experience causes breeding success and birds that breed successfully get more nesting experience.

(5) A causes C and C causes B
Nesting experience causes blackbirds to become better at avoiding predators and avoiding predators results in greater breeding success.

The author argues that the true relationship here is option (1) - A causes B. Do any of the answer choices suggest otherwise?

(A) is irrelevant. We are only interested in blackbirds, not other species;

(B) is correct because it demonstrates that there is a confounding variable - blackbird's age, the older the bird is, the higher their reproductive rate/ breeding success is and the more skilled they are at building nests;

(C) is wrong because it strengthens our assumption;

(D) is irrelevant because the author argues his conclusion is true even when controlling for size and strengths. -"[t]his cannot be a mere matter of size and strength," meaning we are not interested in the breeding success of weak v. strong blackbirds but rather birds with more nesting v. less nesting experience;

(E) is irrelevant and has no impact on the validity of the conclusion.

I know correlation and causation is a tricky topic, but it sounds like you are on the right track. Let me know if this makes sense and if you have any other questions.