[R] help on ANOVA, please

netrunner giovanna.varni at gmail.com
Tue Feb 22 17:04:53 CET 2011


Dear all,
I need to perform a two-way unbalanced ANOVA on my data set. The unbalancing
is due to missing values (NA) for some subjects. My data set concerns the
evaluation of 4 different prototypes tested by subjects three times and it 
is organized as follows (column by column):

number of observation    number of subjects   type of prototype  replication 
score

replication is from 1 to 3 and score is the score that each subject gave for
each replication for each prototype. 

I know that I cannot use aov(), so I used the function lm in the following
way:

a.lm<-lm(Score~(Prototype*Replication),na.action=na.omit,data=a.ex)
Anova(a.lm,type=c("II"))

the output was:
Anova Table (Type II tests)

Response: Score
                                 Sum Sq  Df F value    Pr(>F)    
Prototype                    216.61   3 10.1598 1.816e-06 ***
Replication                  49.24   1  6.9287  0.008803 ** 
Prototype:Replication    2.91   3  0.1366  0.938133    
Residuals  2906.69 409 

I used these in the right way? I extensively read previous post on Anova(),
lm () and so on, but I am confused.
Then, I would like to verify the normality of residuals and the homogeneity
of variance.  My idea was to apply the shapiro test on the residuals and the
bartlett or the levene test.

Please, someone can provide me with some suggestion?

thank you for your attention

netrunner

-- 
View this message in context: http://r.789695.n4.nabble.com/help-on-ANOVA-please-tp3319467p3319467.html
Sent from the R help mailing list archive at Nabble.com.



More information about the R-help mailing list