0:00:00.190,0:00:02.660
- [Instructor] We are asked[br]what is the critical value,

0:00:02.660,0:00:05.790
t star or t asterisk, for constructing

0:00:05.790,0:00:09.310
a 98% confidence interval for a mean

0:00:09.310,0:00:14.170
from a sample size of n is[br]equal to 15 observations?

0:00:14.170,0:00:16.630
So just as a reminder[br]of what's going on here,

0:00:16.630,0:00:18.370
you have some population.

0:00:18.370,0:00:19.580
There's a parameter here,

0:00:19.580,0:00:21.410
let's say it's the population mean.

0:00:21.410,0:00:24.970
We do not know what this[br]is, so we take a sample.

0:00:24.970,0:00:27.580
Here we're going to take a sample of 15,

0:00:27.580,0:00:30.990
so n is equal to 15, and from that sample

0:00:30.990,0:00:33.200
we can calculate a sample mean.

0:00:33.200,0:00:36.340
But we also want to construct[br]a 98% confidence interval

0:00:36.340,0:00:38.010
about that sample mean.

0:00:38.010,0:00:39.880
So we're going to go take that sample mean

0:00:39.880,0:00:42.800
and we're going to go plus or[br]minus some margin of error.

0:00:42.800,0:00:44.510
Now in other videos we have talked about

0:00:44.510,0:00:46.690
that we want to use[br]the t distribution here

0:00:46.690,0:00:49.850
because we don't want to[br]underestimate the margin of error,

0:00:49.850,0:00:52.665
so it's going to be t star times

0:00:52.665,0:00:56.200
the sample standard deviation divided by

0:00:56.200,0:00:59.340
the square root of our sample[br]size, which in this case

0:00:59.340,0:01:02.320
is going to be 15, so[br]the square root of n.

0:01:02.320,0:01:03.570
What they're asking us is

0:01:03.570,0:01:05.760
what is the appropriate critical value?

0:01:05.760,0:01:10.370
What is the t star that we[br]should use in this situation?

0:01:10.370,0:01:13.620
We're about to look at, I[br]guess we call it a t table

0:01:13.620,0:01:16.890
instead of a z table, but[br]the key thing to realize

0:01:16.890,0:01:19.670
is there's one extra variable[br]to take into consideration

0:01:19.670,0:01:23.090
when we're looking up the[br]appropriate critical value

0:01:23.090,0:01:27.340
on a t table, and that's this[br]notion of degree of freedom.

0:01:27.340,0:01:29.670
Sometimes it's abbreviated df.

0:01:29.670,0:01:31.970
I'm not going in depth[br]on degrees of freedom.

0:01:31.970,0:01:34.720
It's actually a pretty deep concept,

0:01:34.720,0:01:36.880
but it's this idea that you[br]actually have a different

0:01:36.880,0:01:40.670
t distribution depending on[br]the different sample sizes,

0:01:40.670,0:01:42.540
depending on the degrees of freedom,

0:01:42.540,0:01:44.990
and your degree of freedom is going to be

0:01:44.990,0:01:47.490
your sample size minus one.

0:01:47.490,0:01:50.060
In this situation, our degree[br]of freedom is going to be

0:01:50.060,0:01:54.100
15 minus one, so in this[br]situation our degree of freedom

0:01:54.100,0:01:56.160
is going to be equal to 14.

0:01:56.160,0:01:58.450
This isn't the first time[br]that we have seen this.

0:01:58.450,0:02:00.530
We talked a little bit[br]about degrees of freedom

0:02:00.530,0:02:03.470
when we first talked about[br]sample standard deviations

0:02:03.470,0:02:05.210
and how to have an unbiased estimate

0:02:05.210,0:02:07.390
for the population standard deviation.

0:02:07.390,0:02:09.840
In future videos we'll go into[br]more advanced conversations

0:02:09.840,0:02:12.080
about degrees of freedom,[br]but for the purposes

0:02:12.080,0:02:14.640
of this example, you need to know that

0:02:14.640,0:02:17.380
when you're looking at the t distribution

0:02:17.380,0:02:19.640
for a given degree of freedom,[br]your degree of freedom

0:02:19.640,0:02:21.490
is based on the sample[br]size and it's going to be

0:02:21.490,0:02:23.760
your sample size minus one[br]when we're thinking about

0:02:23.760,0:02:26.280
a confidence interval for your mean.

0:02:26.280,0:02:29.080
Now let's look at the t table.

0:02:29.080,0:02:32.005
We want a 98% confidence interval

0:02:32.005,0:02:35.893
and we want a degree of freedom of 14.

0:02:36.890,0:02:40.010
Let's get our t table out, and I actually

0:02:40.010,0:02:41.940
copied and pasted this[br]bottom part and moved it up

0:02:41.940,0:02:43.650
so you could see the whole thing here.

0:02:43.650,0:02:45.300
What's useful about this t table

0:02:45.300,0:02:47.300
is they actually give[br]our confidence levels

0:02:47.300,0:02:50.450
right over here, so if you[br]want a confidence level of 98%,

0:02:50.450,0:02:53.020
you're going to look at this column,

0:02:53.020,0:02:55.580
you're going to look at[br]this column right over here.

0:02:55.580,0:02:59.160
Another way of thinking about[br]a confidence level of 98%,

0:02:59.160,0:03:02.620
if you have a confidence level of 98%,

0:03:02.620,0:03:07.230
that means you're leaving 1% unfilled in

0:03:07.230,0:03:09.680
at either end of the[br]tail, so if you're looking

0:03:09.680,0:03:12.780
at your t distribution,[br]everything up to and including

0:03:12.780,0:03:16.620
that top 1%, you would[br]look for a tail probability

0:03:16.620,0:03:21.450
of 0.01, which is, you[br]can't see right over there.

0:03:21.450,0:03:23.040
Let me do it in a slightly brighter color,

0:03:23.040,0:03:25.820
which would be that tail[br]probability to the right.

0:03:25.820,0:03:27.650
Either way, we're in this[br]column right over here.

0:03:27.650,0:03:30.080
We have a confidence level of 98%.

0:03:30.080,0:03:32.330
Remember, our degrees of freedom,

0:03:32.330,0:03:37.193
our degree of freedom here,[br]we have 14 degrees of freedom,

0:03:37.193,0:03:41.613
so we'll look at this row right over here.

0:03:41.613,0:03:43.210
So there you have it.

0:03:43.210,0:03:46.333
This is our critical t value, 2.624.

0:03:48.750,0:03:51.200
So let's just go back here.

0:03:51.200,0:03:56.200
2.264 is this choice right[br]over here, and we're done.