Whoops, sorry about that. She was doing the experiment to test the effect of the background color on the completion rate. White was the worst at ~81% completion and orange the best at ~88%, but with a 95% confidence interval of +/- ~10% (at best) all she could really conclude was that the sample size was too small. A "could be something, could be nothing" result. Thanks much to everyone who participated. ^.^
