00:01
This is problem number 94 out of section 9 .3.
00:07
They're asking us to look at a study with 5 ,000 people in it, and the results were statistically significant at a 5 % level.
00:15
And why might this not be particularly important? well, we always want to sample at least a size of 30, so we can use a central limit theorem.
00:25
And we're always saying that bigger sample sizes are better, but not quite when we are doing hypothesis testing.
00:34
When we do hypothesis testing, we've sort of got to think of standard error.
00:40
Now think of the formula for standard error here.
00:43
Standard error is the sample standard deviation divided by the square root of a sample size.
00:53
If this sample size gets really large, say 5 ,000, that means, when i divide by that number, let's see square root of 5 ,000, that means my denominator is 70 .7...