BrandX {pda} | R Documentation |
These are results of a detergent brand preference study, with
respondents classified by temperature and softness used during their
wash cycle. Venables and Ripley call Brand
the response
factor and M.user
, Temp
and Soft
stimulus
factors. Thus they are most interested in Brand
and in
interactions that involve Brand
.
data(BrandX)
Ries and Smith (1963) The use of chi-square for preference testing in multidimensional problems. Chemical Engineering Progress 59, 39-43.
# Four-way Contingency Table data( BrandX ) # fit poisson model BrandX.fit <- glm( Fr ~ M.user*Temp*Soft+Brand, family = poisson, data = BrandX ) anova( BrandX.fit, test = "Chisq" ) drop1( BrandX.fit, formula( BrandX.fit), test = "Chisq" ) BrandX.step <- step(BrandX.fit, list( lower = formula( BrandX.fit ), upper = ~.^3 ), scale = 1, trace = FALSE) BrandX.step$anova anova( BrandX.step, test = "Chisq" ) BrandX.mod <- glm( terms( Fr ~ M.user*Temp*Soft + Brand*M.user*Temp, keep.order = TRUE ), family = poisson, data = BrandX ) summary(BrandX.mod, correlation = FALSE, test = "Chisq" )