0:00 what is estimation well there's no way
0:03 you can sample the entire population the
0:06 only thing you can do is get a sample of
0:08 it from that sample hopefully the sample
0:11 is large enough because we've decided
0:12 what our outcome is going to be we can
0:14 generate an estimate of how much power
0:17 or how much sample size we're going to
0:19 need to detect that difference go out
0:21 and sample it generate parameters from
0:25 that what's a parameter parameter of a
0:34 model mean standard error those are
0:44 parameters that come and play a role in
0:46 defining the model and then an
0:50 estimation once you get the model if
0:52 it's right which is a big if sometimes
0:55 but large enough sample size do you guys
0:59 remember the central limit theorem from
1:01 statistics central limit theorem what is
1:06 that basically the central limit theorem
1:10 says no matter what you start with if
1:12 you get a large enough sample size with
1:14 repeated sampling you're gonna get a
1:18 normal distribution ultimately
1:20 everything will collapse to a normal
1:23 bell shaped curve with enough sample
1:25 your sample may not be large enough to
1:28 do that so you might have to use an
1:31 alternate distribution like chi-square
1:34 distribution or like an F distribution
1:36 or a Pareto distribution there are many
1:40 different types of distributions a T
1:42 distribution ultimately with enough
1:44 sample size in the correct model they
1:46 will converge to being a normal the
1:49 normal allows you to make
1:50 generalizations about the larger
1:52 population right larger population for
1:55 more or less is going to be normal so if
1:58 your sample has the same distribution as
2:00 the population your sample is a good
2:02 sample and allows you to make estimation
2:04 on that population now hence the crystal
2:07 ball kind of feeling statistics allows
2:09 you to get a feeling from what the gym
2:11 population would have
