Part 5: Unbalanced 2-Way (Fixed Effects) ANOVA

Dataset: celebrityendorse_attention.csv

Source: P.-S. Wei, H.-P. Lu (2013). "An examination of the Celebrity

Endorsements and Online Customer Reviews Influence Female Consumers'

Shopping Behavior," Computers in Human Behavior, Vol. 29, pp. 193-201.

Description: Experiment with 2-factors: Product (1=Shoes, 2=Ink Toner)

and Endorser (1=Celebrity, 2=Online customer). Completely Randomized

(Between Subjects) Design. Data simulated to match trt means and SDs.

Response is subjects' attention score (average of 3 1-5 responses).

Variable names:

trt_group /* 1=shoes/celeb, 2=shoes/online, 3=toner/celeb, 4=toner/online

product /* 1= shoes, 2=ink toner */

endorser /* 1=celebrity, 2=online customer */

shoe /* 1=shoe, 0=toner */

celebrity /* 1=celebrity, 0=online customer */

attention /* response variable */

• Fit the 2-Way ANOVA model as a regression model with X₁ = 1 if shoes, -1 if Toner; and X₂ = 1 if Celebrity, -1 if online customer. Test for interaction, and main effects.
• Use the aov function in R. Test for main interaction and main effects. (use the contrasts=c("contr.sum", "contr.poly") option)
• Compare the results.