00:01
So in this question, it's a little bit vague, but we're asked about bootstrapping.
00:03
It says, if we sample students from a single student residence hall, does it have an impact on the uncertainty, the range of the compatibility interval computed in bootstrapping? so just to discuss quickly what bootstrapping is, bootstrapping is when it's somewhat of a simpler way to compute confidence intervals if you have a sample, but maybe you are not quite sure about the right assumptions to use, or the right distribution to be using for your confidence intervals.
00:35
I mean, easier thing is to just start with your sample and then repeatedly sample smaller samples from it to do various inference and compute various statistics from that and compute confidence intervals.
00:50
The question here is a little vague, because it's asking, does sampling from a single student residence hall have an impact on our confidence intervals? i would say probably to make this more well -defined, you really need to ask, as opposed to what? are we talking about a larger population of, i guess, all students across all residence halls? in that case, it certainly will.
01:18
If you imagine the thing that you're trying to estimate is something that people in this residence hall have specific value of, then yeah, if you sample students from a single residence hall, you're going to end up, they're all going to look pretty similar.
01:38
And so your confidence intervals, the standard deviations you measure will be smaller, so your confidence intervals will be smaller.
01:45
Alternatively, you'd want to sample from the full population...