A retailer is interested in determining which ads are most associated with drawing consumers to its website. It creates three types of ads, coded 1, 2, and 3. The retaller believes that there may a relationship between the type of ad and the website on which the ad is
seen, so the ads will be put on three types of sites: news (coded level 1), sports (coded level 2), and entertainment (coded level 3). Each of the treatments will be randomly assigned to websites several times over the course of a week. The response to be measured is
the average number of visits to the retaller's site ("clicks") for each of the treatment combinations. The treatment averages are summarized in the table below.
Ad 1 Ad 2 Ad 3
Site 1 26.3 25.2 25.6
Site 2 24.8 25.3 26.1
Site 3 27.0 26.7 25.3
(a) Which of the following correctly interprets the interaction between site and ad type?
Since the lines are not parallel, there is no interaction between site and ad type.
Since the lines are not parallel, there is likely an interaction between site and ad type.
Since the lines are not parallel, we should examine only the main effects for ad type and for site type.
It is reasonable to ignore the interaction for this study since there is no clear pattern in the interaction
plots.
(b) The largest treatment mean occurs in the 3 level of ad type and the 1 level of website type.
(c) For website level 1, ad type level 3 had the highest group mean clicks.