[R] Is manova() correct for this analysis?

angelo.arcadi at virgilio.it angelo.arcadi at virgilio.it
Tue Nov 24 00:58:59 CET 2015

Dear list members,
I have to perform the following analysis but I do not know which function in R 
must be used. My guess is to use manova().

During an experiment I presented participants with some sound stimuli and I asked them to 
modify two parameters of the sound (Centroid and Sound_Level_Peak) to reach a given 
experimental goal. The initial sounds (preset sounds) had a value for those two parameters. 
What I am interested in is whether participants' modifications of the two parameters of the 
sound stimuli resulted in values actually different from the initial values of the parameters
of the preset sounds.

To give an idea, some rows of my data set are the following:

> head(scrd) 
Stimulus_Type   Centroid        Sound_Level_Peak    Preset
Stimulus_A      1960.2          -20.963                           no
Stimulus_A      5317.2          -42.741                           no
Stimulus_B      11256.0        -16.480                           no
Stimulus_B      9560.3          -19.682                           no
Stimulus_A      1900.2          -18.63                             yes
Stimulus_A      5617.6          -44.41                             yes
Stimulus_B      12056.0        -17.80                             yes
Stimulus_B      8960.5          -21.82                             yes

This is the analysis I performed with manova():

> fit <- manova(cbind(Centroid,Sound_Level_Peak)~ Stimulus_Type*Preset, data=scrd)
> summary(fit, test="Pillai")
                       Df  Pillai approx F num Df den Df  Pr(>F)    
Stimulus_Type          11 0.91888  106.629     22   2760 < 2e-16 ***
Preset                  1 0.00343    2.371      2   1379 0.09378 .  
Stimulus_Type:Preset   11 0.01155    0.729     22   2760 0.81348    
Residuals            1380                                           
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

If I am not wrong, these results say that for each stimulus type there is no difference 
between the patterns of the two parameters in the preset and modified conditions.

Can anyone please tell me if I am correct or suggest how to perform in R such an analysis?

Thanks in advance



	[[alternative HTML version deleted]]

More information about the R-help mailing list