Bonferroni after MANCOVA

[Guillaume Geri]

Hi everybody
I'm using MANCOVA to evaluate the adjusted role of several epidemiological variables with health related quality of life SF-36 questionnaire (which has 8 different continuous dimensions).
I performed MANCOVA as follows :
manova PF_trans RP_trans BP_trans GH_trans std_PHYS=c.age sexe rythmechoc c.dureerea lieu_recode rea_temoin temoin_present pci hypoT
and it ran well.
I would like to apply Bonferroni test after that, using manovatest. However, I need matname (for the option test(matname)), that is, if i have correctly understood, the matrix generated by MANCOVA. But where is this matrix. Do I have to build it myself ? If so, any help ?
Thanks in advance
Guillaume

[Carlo Lazzaro]

ggeri (as per FAQ 6, please take a look at how to re-register in line with the preferred requirements of this forum. Thanks):
I would point you out to the Example #2 under -manova postestimation- entry in Stata 13.1 .pdf manual.

[Marcos Almeida]

On a slightly different note, I fear that, may the questionnaires be somewhat correlated (as I expect, since they seem to reflect QOL), maybe a discriminant analysis would better "tell them apart" in terms of different dimensions, in case they exist. With 5 DVs, the Bonferroni post hoc adjustments for all comparisons may well become too conservative.

[Carlo Lazzaro]

You may also want to take a look at -help pca-

[Guillaume Geri]

Thanks a lot.
I've carefully read the Stata manual which is very clear. However, how can I decide which factors I have to use in the matrix ? I wonder it's not supposed to be randomly chosen !
Excuse me for my beginner's question..

[Carlo Lazzaro]

ggeri (as per FAQ 6, please take a look at how to re-register in line with the preferred requirements of this forum. Thanks):
you may find guidance on what you're after in Example 1: Principal component analysis of audiometric data in -pca- entry in Stata 13.1 .pdf manual.