You are on page 1of 6

1

Using SPSS: One-Sample Hypothesis Tests


Create/Load your data: (last page)
Hypothesis Test about a Mean
unknown t-test - SPSS is great at this!
1. Open the sample data (below left).
Here the sample mean is 38.6 and the sample standard deviation is 8.5.
2. Example: Suppose we want to test the claim that the population mean is larger than 35 (Or the
mean score of 38.6 is signicantly more than 35).
H
0
: = 35 H
1
: > 35
3. In the Statistics Viewer choose Analyze Compare Means One-Sample T Test . . .
4. This opens another dialogue box.
2
5. The results of this test are displayed in the Statistics Viewer.
6. Here, the two-tailed P-value is 0.123. Since we are conducting a one-tailed test, the P-value is
0.123/2 = 0.0615. Whether or not you reject the null hypothesis depends on the level of signicance
you have. I.e., How willing are you to make a mistake?
3
Hypothesis Test about a Proportion
Binomial Test - A lot more complicated than the previous test!
1. Open the sample data (below left).
Here the level of education is ranked from 1 (did not nish high school) to 5 (graduate degree). In
this data set, 21.7% of the people surveyed had an education score of 1.
2. Example: Suppose we want to test the claim that the population proportion is 0.20.
H
0
: p = 0.2 H
1
: p = .2
3. In the Statistics Viewer choose Analyze Nonparametric Tests One-Sample . . .
4. A dialogue box opens. Under the Objective tab, choose Customize analysis.
4
5. Under the Fields tab, transfer the data to the Test Field.
6. Under the Settings tab, check Compare observed ..., the click on options.
7. This opens another dialogue box. Set the Hypothesized proportion (0.2) and the cut point to 1.
Click OK and then Run .
5
8. The results of this test are displayed in the Statistics Viewer.
9. Here, the two-tailed P-value is 0.000 (rounded to 3 decimals). Clearly we reject the null hypothesis
that the proportion of level 1 education is not 20%. Because the sample percentage is 21.7% we
can reasonably conclude that the population proportion is greater than 20%.
6
Create/Load your data:
1. Create your own data from an empty data le or open an existing data le.
2. When you open an existing data le, you will be prompted to answer some questions. For example,
if you open an Excel le it will ask whether or not the rst row contains labels.
3. When your data le is open, click the Variable View tab at the bottom of the data le and make
sure all your variables are dened appropriately. The most important distinctions are
TYPE: Numeric (quantitative) or String (qualitative).
Measure: Scale (ratio), ordinal, or nominal (qualitative data).

You might also like