Why is it essential to avoid inferring causation from a...

Why is it essential to avoid inferring causation from a correlation? A. Because correlations are often affected by experimental treatments. B. Because correlations only work in controlled laboratory settings. C. Because other unknown variables could be influencing both correlated variables. D. None of the above.

Transcript

00:01 So we're looking at bivariate correlations.

00:02 So we have two variables and we're looking at how they're related to each other.

00:06 And we say that correlation does not equal causation.

00:13 Why do we say that? ok, so first of all, what can we say about bivariate correlations? we have these two variables, x and y, and we want to know how reliably change.

00:25 Change, so how reliably, change in one variable predicts change in the other.

00:44 And that is a measure of the strength of the correlation.

00:48 The correlation could be positive, which means as one variable goes up, so does the other.

00:52 If it goes down, so does the other.

00:54 It could be negative, as one goes up, the other goes down.

00:56 But either way, we're looking at how reliably we can predict change based on the other variable.

01:01 So let's look at our options here.

01:04 We have four.

01:05 I'm going to start by ruling out d.

01:08 Correlations are a great form of analysis.

01:11 It's really helpful to know how well you can predict one variable using another one that might be easier to measure or might actually come before the variable of interest.

01:22 So we definitely use correlations.

01:24 Is it difficult to establish if they are significantly related? no, that's fine.

01:30 You can definitely do that.

01:31 We often do something called regression analysis to do this...

Question

Please give Ace some feedback