In Exercises 50 and 52, we see the number of advertised brand names people recalled immediately after watching TV shows and 24 hours later. Strangely enough, it appears that they remembered more about the ads the next day. Should we conclude this is true in general about people’s memory of TV ads?
a) Suppose one analyst conducts a two-sample hypothesis test to see if memory of brands advertised during violent TV shows is higher 24 hours later. The P-value is 0.00013. What might she conclude?
b) Explain why her procedure was inappropriate. Which of the assumptions for inference was violated?
c) How might the design of this experiment have tainted these results?
d) Suggest a design that could compare immediate brand name recall with recall one day later.