Note that the motor was nominally rated at 50 Hz A/C to run 1440 rpm. We adopt
the null hypothesis and must say there is no proof that paddlewheel A performed
better than B, although the standard errors of Table 4 do not quite overlap. The
null hypothesis is strictly adopted because t-tests of Table 5 show no significant
difference. Conversely, all indications are that there certainly is a significant
improvement in SOTR comparing paddlewheel C to either of paddlewheels A or B
operating at full speed.