I have 2 dependent variables (DVs) each of whose score may be influenced by the set of 7 independent variables (IVs). DVs are continuous, while the set of IVs consists of a mix of continuous and binary coded variables. (In code below continuous variables are written in upper case letters and binary variables in lower case letters.)
The aim of the study is to uncover how these DVs are influenced by IVs variables. I proposed the following multivariate multiple regression (MMR) model:
my.model <- lm(cbind(A, B) ~ c + d + e + f + g + H + I)
To interpret the results I call two statements:
summary(manova(my.model))
Manova(my.model)
Outputs from both calls are pasted below and are significantly different. Can somebody please explain which statement among the two should be picked to properly summarize the results of MMR, and why? Any suggestion would be greatly appreciated.
Output using summary(manova(my.model))
statement:
> summary(manova(my.model))
Df Pillai approx F num Df den Df Pr(>F)
c 1 0.105295 5.8255 2 99 0.004057 **
d 1 0.085131 4.6061 2 99 0.012225 *
e 1 0.007886 0.3935 2 99 0.675773
f 1 0.036121 1.8550 2 99 0.161854
g 1 0.002103 0.1043 2 99 0.901049
H 1 0.228766 14.6828 2 99 2.605e-06 ***
I 1 0.011752 0.5887 2 99 0.556999
Residuals 100
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Output using Manova(my.model)
statement:
> library(car)
> Manova(my.model)
Type II MANOVA Tests: Pillai test statistic
Df test stat approx F num Df den Df Pr(>F)
c 1 0.030928 1.5798 2 99 0.21117
d 1 0.079422 4.2706 2 99 0.01663 *
e 1 0.003067 0.1523 2 99 0.85893
f 1 0.029812 1.5210 2 99 0.22355
g 1 0.004331 0.2153 2 99 0.80668
H 1 0.229303 14.7276 2 99 2.516e-06 ***
I 1 0.011752 0.5887 2 99 0.55700
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1