Return to PSYC 4650 Opening Page
Creating Comparison Groups from More than One Question (3 22 12)
Background and Rationale
Sometime the best design for your study will involve one target group and
perhaps two or three control groups.
Extra control groups may help you to better pinpoint the specific effects
of the variable you are studying.
·
In order to perform the statistical tests, you will probably need to create a
new variable which includes all your
groups. For example, a new variable
“Diagnosis” might include -- Group1 (ADHD, but not LD), Group 2 (LD but not
ADHD), and Group 3 (Neither ADHD nor LD). This new
Diagnosis variable will be (or one
of) the independent variable(s) for your statistical tests.
·
This can be accomplished by using the
Data → Select Cases → If method.
The first step
is to create your groups, randomizing and fixing group sizes as appropriate
(group sizes should be approximately equal). When you have 2 -3 files containing
all the groups you need for your analysis, you will need to merge them.
1. Open one of the files to be merged,
then –
2. Data → Merge File → Add Cases →
enter the name of one of the other files to be merged
→ Save As → and give this new file a
name.
3. Open the new file you just created
4. Data → Merge File → Add Cases →
enter the name of the last file to be merged
→ Save As → give this file an
appropriate name, e.g. “finalDataSet”, and
save it in your H drive.
You now have a data set containing all of your controlled groups for your Final
project.
Creating a new variable to separate your groups
1. Create
a new variable in the “Variable View”
of SPSS by typing a new name in the “variable name” column, for example,
“Diagnosis.”
Your data set should already be grouped up by the way you merged your files. If
not:
Data → Sort Cases →
Sort By → ADHD, SEX, or whatever
variable groups your data appropriately in the order you need it, i.e. Group 1,
Group 2, etc. Now your data will be sorted and ordered together.
2.
Find your first group, e.g. the main group you are investigating - “variable1 =
1 & variable2 = 2”.
3.
Enter a “1” in your “Diagnosis” variable column in the
Variable View
4. Copy
the “1”
- Click on the first cell and then
Shift+Click the cell of the last member of
Group 1, then
paste the “1” into all of the
columns.
Repeat this process with your other two groups, i.e. “variable1 = 2 & variable2
= 1” becomes Group 2 of “Diagnosis”
and “variable1 = 2 & variable2 = 2” as
Group 3. Run a Frequency
Distribution to ensure your work is correct.
Now you have your final project data set with a built-in independent variable:
Diagnosis.
You will probably use this new
Diagnosis variable for statistical tests such as ANOVA.