I have two groups receiving two alternate drugs. The comparison of groups revealed us no differenceof outcomes.
However I need age and birth weight adjustment of two groups and adjusted p values.
Is it possible with logistic regression or some other analysis? However, I don't have any specific outcome to look for predictors. Just searched the differences of several outcome parameters. And I don't need to find a predictor. Just want to compare two groups adjusted by age, etc. Is it possible?
I need an advice for SPSS.
Thanks
Sounds like an MANCOVA would do the job, if you got several dependent variables. Or alternatively run an ANCOVA for each dependent variable and correct for multiple comparisons by adjusting the alpha-level (this is more transparent than reporting an adjusted p-value; that is, report the p-values but state that you are only going to interpret p-values below .0x as significant).
To do a MANCOVA in SPSS go to Analyze -> General linear model -> Multivariate
What type is your outcome, continuous? binary? In order to compare two groups, you need an outcome(or outcomes). You should fix your outcome (actually your hypothesis tells this). Then the type of analysis is determined and you adjust by whatever variables you wish.
Can you be more specific?
Sounds like an MANCOVA would do the job, if you got several dependent variables. Or alternatively run an ANCOVA for each dependent variable and correct for multiple comparisons by adjusting the alpha-level (this is more transparent than reporting an adjusted p-value; that is, report the p-values but state that you are only going to interpret p-values below .0x as significant).
To do a MANCOVA in SPSS go to Analyze -> General linear model -> Multivariate
ancova would be appropriate, for a single outcome, mancova if multiple, Age and birth weight should be added as your covariates. The options/plots menus will enable you to compute estimated marginal means, produce a line plot to supplement your significance test, and, by selecting parameter estimates, get a regression style output. Alternatively, if a single DV, just run using linear regression, enter Age and birth weight in first block of predictors, then yuour grouping var in your second - that way you get an r-squared change stat i.e. effect size for your group effect.
I think that the type of statistical analysis depends on type of your outcome (i.e. binary or continuous).
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS