This is probably a very simple question for the statistical minds. I have three columns.
Column 1 - group 1 or group 2 allocation
Column 2 - baseline variable1 (numerical)
Column 3 - 12 month variable1 (numerical)
I have similar analysis to perform for 16 variables. Plus I am also thinking of adding Column 4 - 24 month variable 1 (numerical).
The question is - what statistical test can I use using SPSS to check whether group 1 had a statistical difference between baseline and 12 months and whether group 2 had a statistical difference between baseline and 12 months.
One option is that I load SPSS with only 1 group and do a paired sample t-test.
alternatively, I get the means and SD for all the variables and perform a summary independent sample t-test.
There must be a simpler way to do this? Can anyone help?
What test do I need for baseline, 12 months and 24 months comparison between 2 groups and also within groups (ie group 1 [0-12-24M], and group 2 [0-12-24M])
many thanks.
Habib
Hi Habib Khan ,
To start, I guess the test you are looking for is a two-way mixed ANOVA.
Go to 'Analize' --> 'General Linear Model' --> 'Reapeated Measures'
Provide a name for your within-subject factor ej: time
Define the number of levels (ej., the number the test is repeated) ej: 3 (for month_1, month_12, and month_24).
Click 'define'
In the within-subject variable window (factor 1):
Move month_1 to _?_(1)
Move month_12 to _?_(2)
Move month_24 to _?_(3)
Move you group variable to the between-subjects factor(s).
Click OK
The output will let you know whether there is a significant effect for time, group, or a significant interaction between them.
Then you will want to determine simple main effects, this is, the difference 1) between groups at each category of time and 2) vice versa.
For the first one, running these comparisons is the same as running three separate between-subjects ANOVAs (i.e., three separate one-way ANOVAs).
For this, click Click Analyze --> General Linear Model --> Univariate... on the top menu.
In the Univariate dialogue box transfer the first category of the within-subjects factor, month_1, into the Dependent Variable: box and then transfer the between-subjects factor, group, into the Fixed Factor(s)
Repeat the procedure above for the other category(es) of the within-subject factor. (ej month_12 and month_24, etc).
For the second one, you will need to determine Simple main effect for time, thus, there will be three omnibus statistical tests (simple main effects). Performing these comparisons is the same as running three separate within-subjects ANOVAs (repeated measures ANOVAs). For this, I pre-split (in the “Data” --> “Split file” according to the “group”, select “organize output by groups”, and run repeated measures in the Analyze --> General Linear Model --> Repeated measures menu but remove “group” from the repeated-measures factor.
If there is a significant interaction, you have to check pairwise comparisons to determine where in time and group the differences occur (usually with Bonferroni corrections).
Let me know if it works, and provides meaningful results.
Regards,
Ricardo
- Badges
- Science topic
- Similar topics
- Psychology
- Cognitive Science
- Mind