|
|
Dear listmates,
This problem stems from my wish to create a bar graph based on the performance of 5 groups on two tests. I want to use standard errors in the graph as well, and I do not know how to get adjusted standard errors based on my covariate (sex).
My understanding is that the error bars should overlap if two groups are not significantly different, but mine presently do not overlap though my results suggest there is a lack of significance between two groups (I am using Bonferronni adjustment) - do I need error bars to account for this adjustment as well?
Excuse my ignorance. Here is my syntax. Any help would be deeply appreciated! Thanks!
GLM
F_words Animals BY mcivsncgroup_ WITH Sex
/WSFACTOR = test 2 Polynomial
/METHOD = SSTYPE(3)
/EMMEANS = TABLES(mcivsncgroup_) COMPARE ADJ(LSD)
/EMMEANS = TABLES(test) COMPARE ADJ(LSD)
/EMMEANS = TABLES(mcivsncgroup_*test)
/EMMEANS = TABLES(mcivsncgroup_*test) compare (test) adj(bonferroni)
/EMMEANS = TABLES(mcivsncgroup_*test) compare (mcivsncgroup_) adj(bonferroni)
/PRINT = DESCRIPTIVE ETASQ TEST(MMATRIX)
/CRITERIA = ALPHA(.05)
/WSDESIGN = test
/DESIGN = Sex mcivsncgroup_.
|
Locked
