Tip:
Highlight text to annotate it
X
In this video I will show you how to conduct a test of significance in
SPSS.
Specifically, I will show you how to test whether
the sample mean
for a certain variable is significantly different than a given
value which is still a population of mean.
What I'm interested in doing here is to determine whether
the mean for the post-test variable
for this sample,
it's significantly different then the population mean.
Let's say that
we know from previous administrations of these test that the population
mean was 85.
So, what I want to know is whether this sample mean
is statistically different than that value of 85.
To do that, all I have to do is to go to the
Analyze Menu,
go to Compare Means,
and then choose
One-Sample t Test.
In the window that opens here we have to select the variable that we are
interested in, which in our case is post-test, so we select it,
click on the arrow,
and in this area we have to enter
the population mean
or the value that we are testing
this sample against, so in our case the population mean was 85.
So, we type 85 here
and then click OK.
The output window includes two
tables.
The first table includes just descriptive
statistics, the sample size, which in 20,
the mean for our variable which is 90,
the standard deviation and the standard error
for the post-test variable.
The second table is the one that gives the results of the test of significance.
We can see on the top, here
that the test value was 85,
this was our population mean.
The test statistic t is 2.370.
The test statistic t is 2.370.
It is a positive value,
which means that, the mean for our sample
which was 90, is higher than
the mean of the population, which we know from prior research
that is 85.
The degrees of freedom
are 19.
And the probability, the two-sided probability is .029
.029
.029
So, based on this information because the p value is lower than alpha of .05
which is typically used,
we can reject the no hypothesis, that
states that the sample mean is equal,
to the population mean
of 85.
And we can accept our alternative hypothesis
which is that, the sample mean is different,
then the population mean,
is significantly different than the population mean.
Because our alternative hypothesis was two-sided, then we don't have to
transform this significance level in any way.
If our alternative hypothesis were one-sided,
then we should have divided this significance level by 2 to obtain the p value
then we should have divided this significance level by 2 to obtain the p value