Count-Based Comparison Claims
Ennis, D. M. and Ennis, J. M. (2015). IFPress, 18(2) 3-4.
The National Advertising Division (NAD) is the advertising industry’s self-regulatory body whose mission is to review national advertising for truthfulness and accuracy and foster public confidence in the credibility of advertising. In a recent decision1 , the NAD determined that LG Electronics USA, Inc. (LG) should discontinue a claim that they concluded was unfavorable to Samsung America, Inc. In one version of the ad it was stated that: “In 3D TV Tests, 4 out of 5 People Choose LG Cinema 3D over … Samsung.” The NAD considered a large number of legal and technical issues in preparing this decision. In past NAD decisions, the NAD has held that comparative statements should be justified statistically by accounting for error and they have often relied on 95% confidence or 5% Type I error. As part of Samsung’s challenge, the NAD report referenced the appropriate statistical treatment of count-based comparisons, such as the comparison made by LG. This technical report will consider the procedure for finding the minimum counts needed to report trustworthy, count-based comparison claims.
Figure 1. The count needed (238) to reject a null hypothesis of 0.75 at the 5% level when the sample size is 300.