Scribd
Upload
10 views

8.2V2 Transcript

Uploaded by

hb7
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
10 views

8.2V2 Transcript

Uploaded by

hb7
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

Hypothesis Testing Excel (t-test)

Alright, so in this video, we're going to look at the other next example for hypothesis testing. And
if you just watched the first video on the z-test, we're going to go ahead and work with that
spreadsheet we created and just sort of modify it. We're going to use example 9.16 from the
book, which is on page 484. And still, let's read the problem. For statistics students, believe the
mean score on the first statistics test is 65. A statistics instructor thinks the mean score is higher
than 65. He samples 10 statistics students and obtains the scores: 65, 65, 70, 67, 66, 63, 63,
68, 72, 71. He performs a hypothesis test using a 5% level of significance. The data are
assumed to be from a normal distribution.

Okay, so this is still testing the mean, clearly, right? Since we're looking at finding the mean
score test score. But this is with the population standard deviation unknown, right? And notice if
you read through the problem, we are not given any population standard deviation. Thus, it is
unknown. This is going to be the case where we have sample data. So let's go over there.

All right, your hypotheses still have the same symbol mu because it's a mean test. But if you
think about the instructor's claim, a statistics instructor thinks the mean score is higher than 65.
Right there is the claim of the instructor. That is your alternative hypothesis. Right? And of
course, that translates to the mean being greater than 65. Right, higher than 65 is greater than
65. So, of course, this is a 65 here as well. And the null can stay as equals, or you can make it
less than or equal.

So this is a right-tailed test because of the inequality sign there pointing to the right. Now we are
not able to use the normal distribution because sigma or population standard deviation is
unknown. In order to use the t-distribution, we just need to know that the data are from a normal
distribution originally, and then we can have a small sample size. We can have no population
standard deviation, and we can work just with sample data. But you notice that we are told the
data is from a normal distribution, and I think you need no outliers as well. A lot of times you
don't know that for sure. You can make that assumption and work with the t-test anyway. So
those are the requirements there for the t-distribution to be used. The level of significance here
is the same. It's 5 percent.

And then we go down to our sample statistics. All right, so remember these are actually
calculated, and in the t-test case, we don't have a population standard deviation. We have a
sample standard deviation, so the sample standard deviation will be calculated as well. All right,
well, we need to put in our sample data. So let's go ahead and type those numbers in. And just
delete the empty or the numbers we didn't use. All right, now the sample mean is correct. The
sample size is correct, but we need to have the standard deviation be a calculation here. So
we'll do equals STDEV and then we'll select the sample data. And you can go just as far down
as you need or to make this usable more for later larger data sets and go down really far. So hit
page down a few times. Close parentheses. And you see I went down to 80. So that will cover
you for a sample size of about 60, which is fine, certainly for anything in this course.
All right, so now we have that, and that is a calculation. All right, well, we can approach the
p-value calculation the same way as we can with the normal distribution, and that's because of
Excel's limitations. You notice if I just try to take something like this, right, and replace it with
instead of NORM.DIST, right, the equivalent T.DIST. Yeah, T.DIST. You'll notice that it wants to
know where to draw the line on the distribution, and it wants the degrees of freedom and the
number of tails, and it does not allow you to specify the standard deviation. So what they've
done here is they've only given you the standardized T distribution. Remember, for the normal
distribution, when that was first brought up, there was the standard normal distribution, and it
had the mean of zero and the standard deviation of one. Well, that's what they're giving us here
for the T distribution. So they're only giving us the standard one, and that means that we have to
standardize our sample statistics. We have to standardize them in order to find this picture.

So it's not going to like what I have there. Okay. So the picture we're looking at here is still going
to hold, but we can't actually use that distribution directly. So there's a little shortcut, a middle
step, and it's called the test statistic. And that's what we're going to have here is our test
statistic. Actually, we don't want it there. We still want the standard deviation of the sampling
distribution, I guess. Now we don't. Let's do that. So test statistic. All right, so the test statistic is
basically like a Z-score if you remember that calculation. And remember the Z-score, you take
the individual, and the individual here is going to be the sample mean. And then you subtract
the population mean, and that's going to be the um, this number here, right? And then you
divide by the standard deviation. And so we'll use the sample standard deviation divided by the
square root of the sample size. That's the normal standardized error. Take the standard
deviation and divide by the square of the sample size. That's because of the central limit
theorem. That's your standard deviation for your sampling distribution, right? Okay, and so that's
a, that's a Z-score formula. You have the individual minus the mean divided by the standard
deviation, but we're using the sampling distribution there.

You might also like