00:01
Why is randomization a critical condition in hypothesis testing for comparing two groups? so randomization is where a subject is equally likely to be in the control versus treatment group if you're assigning them.
00:28
Or if you're looking at sampling from a population, all members of a population are equally likely to be sampled.
00:44
So it's all about equal chance.
00:48
The second one is what we usually deal with when we're taking samples.
00:52
Now, why is this a big deal? so there's the kind of intuitive, it makes sense for your sample, and then there's the mathematical reason.
01:01
So first, the kind of intuitive way, is that it balances confounding variables.
01:16
If you want to see if x affects y, but there are loads of other variables running around as well.
01:23
You want to try to balance those other variables.
01:26
You want to make your two groups as similar as possible except for this one thing that you're looking at.
01:32
If you can do that, then any difference between them, you can put down to this one difference.
01:38
So it's making sure that you avoid these confounds...