00:01
All right, we have a bunch of questions about regression here.
00:03
So the first thing we need to remember that is super important is that correlation does not imply causation, right? so that's kind of the point of this first problem here, is that just because two things are correlated does not mean that you can say that one causes the other, okay? so that's the answer to the first one where they say, like, you know, our data suggests that one thing causes another, you can't say that just based on correlation.
00:34
Okay, so you want to look at the one that says that we cannot conclude based on the correlation that one caused to the other.
00:46
Next, if we have a correlation close to zero, so if you think about a correlation being zero, and you think about what that can look like, it can look like a few different things, right? it can look like those random dots.
01:01
It could also look like this, right? but at any rate, if you were to find the slope of the line that goes through that data, the slope would actually be flat.
01:14
Okay? it would look like that.
01:16
Which, that would be the mean of the y's.
01:20
That point would.
01:21
Okay.
01:22
So yes, indeed, when correlation is zero, the best estimate that you have is just the average.
01:28
Because, at least for just like a singular regression line, because we're told that knowing anything about x doesn't tell you anything extra about y, linearly anyways.
01:43
Next, we are asked, how can we think of the mean of the regression line? okay? or, sorry, the intercept...