Answer Happy

4. You measure the change in operating temperature of mobile phone batteries com- pared to the previous year's models for n = 300 popular mobile phone models. Assuming changes are Normal distributed around 0 with a particular variance, you scale your measurements and obtain z-scores. The 5 highest 5 z-scores which you plan to test for significant increases in temperature are: z = {3.59, 3.57, 3.25, 3.14, 2.86). (a) Compute the p-values obtained by independently applying a Wald test to each value in turn. Show your calculations. [3] (b) Assuming a Family Wise Error Rate of 0.05, how many of these do you determine to be statistically significant using a Bonferroni correction for Multiple Testing? [3] (c) Assuming a Family Wise Error Rate of 0.05, how many of these do you termine to be statistically significant, when controlling for Multiple Testing? [5] (d) Now control the False Discovery Rate at q = 0.1. How many values are detemined to be significant? [6]