Thus, in order to have 95 percent confidence that a.; > 0, the observed industry average must be at least 15.21 percentage points above the mean. How often will this happen? The standard deviation CT_v of industry returns has the approximate value 23/0.4 = 7.58. Obviously, the 'typical' industry return does not pass the 15.21 point criterion. Indeed, the 15.21 point
cut-off lies two standard deviations above the mean! Hence, only one in forty industries (FN(2) = 0.977) will exhibit a return large enough to warrant a conclusion that the true industry effects is positive. Put differently, industry returns are such noisy measures of industry effects that only about six of the 242 FTC industries studied could be judged (95 percent confidence) to have positive industry effects