7 Mann–Whitney U SPSS Tutorial
Tải bản đầy đủ - 0trang
1. Type the data into the first two columns in SPSS.
Note: The first column is a vector of brain sizes and the second column assigns the
brain sizes to early (1) or late (2) Homo.
2. Click on the Analyze menu and select Nonparametric, then Legacy Dialogs, and
then click on 2 Independent Samples.
3. The Two Independent-Samples Tests box will appear similar to the picture below.
Click on Homocc to highlight it and click on the arrow next to the Test Variable
List box.
4. Click on the variable for EarlyLate and click on the arrow next to the Grouping
Variable box.
5. Once the two variables are in their respective places in the Test Variable List and
Grouping Variable box, click Define Groups.
6. In the Define Groups box, assign the group numbers in the EarlyLate variable in
the order they appear in the spreadsheet. Then click Continue.
7. SPSS will go back to the Two Independent-Samples Tests screen. Make sure the
Test Type is marked as Mann-Whitney U. Click OK.
8. A separate document will appear. This is referred to as the output.
The Mann–Whitney U value is 8 and the corresponding p-value is low (p = 0.035)
indicating the null hypothesis should be rejected and there are different distributions
between Early and Late Homo.
Concluding Statement
Early and Late Homo display significantly different distributions of cranial capacities (U =
8, p-value = 0.035).
Note: If you want to save the SPSS file with the inserted data as well as the SPSS output
with the results of the statistical analysis performed, then you must save each document
separately (see Chapter 13).
7.8 Wilcoxon Signed-Rank SPSS Tutorial
The following tutorial utilizes some data from the previous Mann–Whitney U test in
SPSS; it was modified for teaching purposes.
The student was further interested in the variations of different reconstructions of
hominid skulls. So he tested the cranial capacities derived from an alternative
reconstruction of each hominid's brain size in his study of early and late Homo, in order
to quantify the disparities between reconstructions.
Formulate a question about the data that can be addressed by performing a
Wilcoxon signed-rank test.
Question: What is the relationship between two different reconstructions of cranial
capacity?
Based on the question, formulate the null and alternative hypotheses that
address the question proposed.
Null Hypothesis (H0): There are no differences in cranial capacities between cranial
reconstructions of the same individuals.
Alternative Hypothesis (H1): There are differences in cranial capacities between
cranial reconstructions of the same individuals.
Now that an appropriate question has been developed, along with a set of testable
hypotheses, you can run the statistical analysis.
To run a Wilcoxon signed-rank test in SPSS, utilize the following tutorial.
Refer to Chapter 13 for getting started and understanding SPSS.
Check all assumptions prior to running the test.
Wilcoxon Signed-Rank Test SPSS Tutorial
1. The Wilcoxon signed-rank test is handled very similar in SPSS to Mann–Whitney U.
The data are loaded slightly differently; the two variables contain cranial capacities for
two different cranial reconstructions of brain size (the original cranial capacity we
used in the prior example and a second estimation on the same crania) and there is no
grouping variable.
2. Click on the Analyze menu and select Nonparametric, then Legacy Dialogs, and
then click on 2 Related Samples.
3. Click on Homocc and the arrow to the left of the Test Pairs box to move Homocc
to the Variable 1 slot.
4. Click on Alternativecc and the same arrow to the left of the Test Pairs box to move
Alternativecc to Variable 2.
5. Click OK.
6. A separate document will appear. This is referred to as the output.