> table(indvar_FODs$categorical_FOD_FODs) ascending descending identity 1683 1432 206 > table(indvar_FODs$Condition_FODs) A B C D 880 880 880 880 > table(indvar_FODs$Language_used_FODs) English German Hungarian Italian Turkish 880 160 160 160 2160 categorical_FOD_FODs is the dependent variable. please apply multinomial logistic regression with interaction effect Treat each level of the Condition_FODs variable (4 levels) and Language_used_FODs variable (5 levels) equally. include the interaction effect of variables in the analysis.
library(nnet) #without interaction effect #m1 <- multinom(categorical_FOD_FODs ~ Condition_FODs + Language_used_FODs, data=indvar_FODs) #summary(m1) #with interaction effect m2 <- multinom(categorical_FOD_FODs ~ Condition_FODs * Language_used_FODs, data=indvar_FODs) summary(m2) #in this case, the interaction effect is not significant #the softmax regression with interaction effect is worse than that without interaction effect #cross validation with bootstrap library(boot) #train the model with all data #without interaction effect model1 <- multinom(categorical_FOD_FODs ~ Condition_FODs + Language_used_FODs, data=indvar_FODs) #with interaction effect model2 <- multinom(categorical_FOD_FODs ~ Condition_FODs * Language_used_
