Which nonparametric test should be used to test the hypothesis that the mean SBPs for the two machines are comparable? 

Hypertension

An instrument that is used in some blood-pressure studies is the random-zero device, in which the zero point of the machine is randomly set with each use and the observer is not aware of the actual level of blood pressure at the time of measurement. This instrument is intended to reduce observer bias. Before using such a machine, it is important to check that readings are, on average, comparable to those of a standard cuff. For this purpose, two measurements were made on 20 children with both the standard cuff and the random-zero machine. The mean systolic blood pressure (SBP) for the two readings for each machine are given in Table 9.13. Suppose observers are reluctant to assume that the distribution of blood pressure is normal. 

Table 9.13: Comparison of mean SBP with the standard cuff vs. the random-zero machine (mm Hg)

 

Mean SBP Mean SBP Person (standard cuff) (random-zero) 79 84 112 99 3 103 92 4 104 103 94 94 106 106 103 97 97 108 88 77 10 113 94 11 98 97 12 103 103 13 105 107 14 117 120 15 94 94 16 88 87 17 101 97 18 98 93 19 91 87 20 105 104