BrandX: Preference for Brand X Detergent

In byandell/pda: Practial Data Analysis for Designed Experiments

Description

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.

Usage

data(BrandX)

Source

Ries and Smith (1963) The use of chi-square for preference testing in multidimensional problems. Chemical Engineering Progress 59, 39-43.

References

Venables and Ripley (1999) Modern Applied Statistics with S-PLUS, 2nd ed., ch. 7. (this is not in the 3rd ed) Cox and Snell (1989) Analysis of Binary Data, Chapman \& Hall.

Examples

# 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" )

byandell/pda documentation built on May 13, 2019, 9:27 a.m.