I have a 2 (group) x 3 (time) analysis performed and there is a difference between the pre values and I'd like to adjust it (using the pre values as covariates). However, on SPSS since I need all values in just one column and I cannot split it up (in the same column I have pre, post1, and post2 values). How to adjust it?
In the figure attached there is a picture of the database.
Hello Dalton,
You will need to reformat your data set. Ideally, the three time/occasion measures should be set in three columns (not one, as you have at present).
Take out the "time" variable/column, and replace the three lines per case (one for time 1, one for time 2, one for time 3) with one row/line per case, in which you have ID, Group, Time 1 CMJ score, Time 2 CMJ score, and Time 3 CMJ score.
Good luck with your work!
Thank you for your response, Professor David Morse .
However, I cannot restructure the data set because the GEE analysis on SPSS requires this exact format. I am wondering if this adjustment is possible in any way on SPSS.
Hello Dalton. First, I have a question: Is ID really ID within group? I.e., is ID=1 in group 2 a different individual than ID=1 in group 1? I suspect it is, and if so, I would recommend computing a new unique ID variable. I think the GENLIN command with GEE may actually require it. You could do something like:
COMPUTE UniqueID = 1000 +ID.
FORMATS UniqueID(F5.0).
Assuming pre = CMJ when time=1, you could then do something like this:
IF Time EQ 2 and uniqueID EQ LAG(uniqueID) PRE = LAG(CMJ).
IF TIME EQ 3 and uniqueID EQ LAG(uniqueID,2) PRE = LAG(CMJ,2).
DESCRIPTIVES PRE.
After doing this, PRE will be missing on rows where Time=1, but filled in where Time = 2 or 3. You should then be ready to proceed with your GENLIN command with GEE. Time will now have only two levels, and PRE can be treated as a covariate. (You could, but don't have to delete the rows with Time=1. You don't have to delete them because with PRE in the model, those rows will be omitted due to listwise deletion.)
HTH.
Dear Professor Bruce Weaver , they are the same individual. I am using this kind of dataset because it is required to perform the GEE analysis on SPSS. The dependent variable has to be placed in one column.
Thank you for your help.
Hi Dalton. If the 6 rows of data for ID=1 are for the same individual, I would change Group to Condition. When you use Group as a variable name, it suggests (to me at least) that you have independent groups.
With Group changed to Cond (for condition), is this what you want the data to look like for ID=1?
ID Cond Time Pre CMJ
1 1 2 32.3 33.5
1 1 3 32.3 31.3
1 2 2 35.6 29.4
1 2 3 35.6 33.4
Sure, I meant condition, I will change it. Also, I will try this dataset, thank you! Bruce Weaver
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS