Hi everyone, having some trouble trying to run linear mixed models on some data I have. I know the problem lies with the study design, but I wanted some advice on whether running 1 or 2 stats models for the below design is appropriate....
I have 3 groups based on exercise-intensity performed (low / medium / high) and 3 timepoints (pre/post/post1hr), my dependent variable is muscle glycogen.
Unfortunately, one of my groups (low) does not have any data available for the final timepoint (post1hr). As such, my LMM terminates in SPSS and I can't perform any comparisons.
Is there any way to get a LMM to run despite this one group missing a whole timepoint's worth of data, or would it be correct to run one LMM for all groups pre & post, and a second for medium & high comparing pre post & post1hr?
Thanks
Dear colleague. I remember when I studied in bachelor level, my lecture tell me how to create the new data from missing data. The formula very complicated. I will looking for my books and tell you imediately. Thanks...
Sir simply take Body Mass Index, Serum free insulin, Insulin resistance in the form of HOMA and S Cholesterol and more reliably serum lipid profile especially HDL !
Good Luck
Wisam
Dear friend
SPSS is a software for those who would like to study statistics in professional way. Mixed linear Model in all other software is called Partial Least Square Fit and simply you have some m number predictors or Independent variables and less number N dependent variables.
PLS fit will show you which predictors of the M group is concerned most with N number dependent variable.
To use SPSS to solve your issue is a waste of time. I strongly advice you to use either MINITAB version 17 or Centurion statographic Texas institute software. They will do some magic work for you! PLS is a synonymy for mixed linear model!
Good Luck
Wisam
Hi you can use some programs solve missing values such as MTDFREML program and WONBAT program
- Badges
- Science topic
- Similar topics
- Correction