Story Name: Food Taste Test
Datafile Name: Food Taste Test
Abstract: Two-way analysis of variance with two levels of each classification can be easily
handled by regression analysis with dummy (0,1) variables for the two levels.
The example involves food taste scores in an experiment in which a coarse and
fine screen for texture and a high and low liquid level were employed in producing
the food. The student should analyze the data using alternative analysis of variance
models (additive vs non-additive) as well as regression models with and without the
interaction (product) variable. The graph below is produced from the output of
the best regression model which can also be obtained through ANOVA if the program
allows specification of the model.
Image:
