Tip:
Highlight text to annotate it
X
Welcome to our helper video on how to run simple T-tests in SPSS. I've already downloaded
the data that I need for this assignment, and I've already opened our program. So, first
thing of course it is asking is for me to import the file which I am going to do now.
Now that I have opened up my data, I see that there is only one column here. If I read my
directions, I know what I am being asked to do is to compare a sample to some known population
mean. The example that I am going to use is height. If we go to the Wikipedia page on
human height, we can scroll down ,and we can find the national average for the U.S in height.
We're going to use this green line here which says that it is 1.763 meters which is 69.5
inches. So let's say that I took a group of Kennesaw students and I wanted to see how
they compared to this national average on height. This is going to be that data on the
55 students I looked at. The first step is to go up to analyze. If you notice there is
not a menu here that says T-tests. However, what we know is that it's here in compare
means, and that we want a one-sampled T-test. But wait a second, note this symbol. That
tells me that my data is nominal. I'm not going to want that. Let's go into variable
view. I see that it is numeric, but its scaled wrong. Instead on nominal, it should be scale.
Now that we fixed that, were going to go back to one-sample T-tests. Note that it's a ruler.
That means its scale data again. So were gonna move that over. Now the test value box is
the box where we put the value we want to run the T-test against. And that again was
five feet nine point five inches which works out in just inches to 69.5. So I'm going to
type that here, and what it's going to do is compare my sample to 69.5. So I'm going
to click O.K. This is our one-sample T-test results. There are many ways that you know
this is a one-sample T-test result. First, it 's clearly labeled one-sample test. It
also lists the test value at the top of the chart. These two charts tell me a lot of very
important information. It tells me how many people were in my sample, what the average
is for that sample, which for this would be the average height, what the standard deviation
is for that error and the standard error of the mean. Further, here we see our T-statistic.
Note that it's negative. We also see our degrees of freedom, and then importantly, our statistical
significance. This data is two groups. Note there is a column that says groups, and the
values in that are one and two which means there are two groups. And then there is a
score. First we want to go in and see that this needs to be numeric, that's numeric,
but the measurement is unknown. So, groups can be ordinal but it's more likely nominal.
We're just naming the groups one and two, whereas their score is actually a scale value.
So now that we have that, what I know is this data represents the score on a test between
one group of students and another group of students. Maybe I'm being asked to compare
whether or not one group did better or worse than another group. Well that's not a one-sample
T-test because we have two samples. So now we have to decide is this within subjects
or between subjects. As there are two groups and no one person can be in both group one
and group two, we know that it's not going to be paired samples. Because of that we are
going to go back up to compare means, and we know that these are independent samples.
If they were paired samples, if you could be in both group one and group two it would
be the paired samples T-test. The independent samples T-test menu looks like this. Note
there is a box that says test variable; this is what we're testing. Score goes here. Then
there is also a grouping variable. For most of your labs your grouping variable will be
the groups that people are in. Know when I move it over it says question mark. This is
because we actually have to define the groups. So what is group one called? Because we just
chose to call them one and two they could be called anything. SPSS wants to make sure.
So group one is literally group one and group two is literally group two. Continue. Now
that our groups are defined, were going to click O.K. Note that there is a lot of difference
between this output and the one-sample test output. This output not only says independent-samples
test above, but we clearly have two groups here and the information for two groups. Also
there is no test value up here. And this is our actual T-test. The first two columns of
an independent T-test does the Levine's test for quality of variances and gives you and
F-ratio and the significance value for that F-ratio. This is not the significance for
the T-test. Depending on whether or not the F-ratio is significant tells us whether or
not we need to read the top line or whether or not we need to read the bottom line. Make
sure you review your notes as to which line you need to read. We have our T-value next,
and then our degrees of freedom and this significance is the significance of that T-value. If these
happened to be paired samples, which they're not, we would go to the paired-samples T-test.
Note it looks very different. We'll have more screen shots on how to do this for the next
lab and there will be plenty of web sites to help you. Let me know if you have any questions.
See you next time.