622 CHAPTER 12 Multisample Inference TABLE 12.31 Mean SBP and difference between machine and human readings at four locations SBP standard cuff SBP machine (mm Hg) SBP machine - SBP standard cuff (mm Hg) (mm Hg) Mean sd n Mean sd Location Mean sd n n ABCD 142.5 21.0 98 142.0 18.1 98 0.5 11.2 98 134.1 22.5 84 133.6 23.2 84 0.5 12.1 147.9 20.3 98 133.9 18.3 98 135.4 16.7 62 128.5 19.0 62 982 84 14.0 11.7 98 6.9 13.6 62 Source: Adapted from the American Heart Association, Hypertension, 2(2), 221-227, 1980. TABLE 12.32 Relationship between clinical status at baseline and follow-up (median follow-up period of 959 days) to mean Mental Function Index at baseline TABLE 12.33 Clinical Status Group Change in DBP among hypertensive patients who receive different kinds of nonpharmacologic therapy Mean change in DBP (baseline - follow-up) (mm Hg) sd change n Baseline Follow-up Mean sd n 1 8.6 6.2 Normal Unchanged 0.04 0.11 27 2 5.3 5.4 Normal Questionably or 0.22 0.17 9 3 4.9 7.0 mildly affected 4 1.1 5.5 2222 Questionably Progressed 0.43 0.35 7 affected Definitely Progressed 0.76 0.58 10 affected Source: Based on the American Journal of Epidemiology, 120(6), 922-935, in diastolic blood pressure (DBP) is noted in these 1984. Suppose 20 hypertensive patients are assigned at random to each of the four groups, and the change patients after a 1-month period. The rou Hypertension Automated blood-pressure measuring devices have ap- peared in many banks, drugstores, and other public places. A study was conducted to assess the comparability of machine readings vs. readings using the standard cuff [13]. Readings were taken using both the machine and the stan- dard cuff at four separate locations. The results are given in Table 12.31. Suppose we want to test whether the mean difference between machine and standard cuff readings is consistent over the four locations (i.e., if the bias is compa- rable over all four locations). 12.9 Is a fixed-effects or a random-effects ANOVA appro- priate here? 12.10 Test whether the mean difference is consistent over all four locations. 12.11 Estimate the proportion of the variance attributable to between-machine vs. within-machine variability.