The text key is a distribution-free key, which is often used when comparing a sample with paired observations, each measured using a different method. Determines whether the levels of measurements differ. The key counts how often there is an increase (+) in the second measurement or decrease (-). Thus, only the sign of the difference of both measurements is considered, hence the name. The test size is the number of pluses (or equivalent the number of minus marks). Under the zero-hypothesis of "no difference in level", the test size comes from a binomial distribution with success rate 1/2.
The cursor key is an improvement of the simple text key. This key also takes into account the size of the difference between both observations. Example
Of the 12 subjects, the (diastolic) blood pressure was measured, both before and after administration of a blood pressure lowering drug. The result was:
Of the 12 subjects, there were 10 cases of decrease and in the other 2 cases of increased blood pressure. Thus, the test magnitude T has the value T = 10 (or alternatively: T = 2). From the B (12,1 / 2) distribution we read for the overrun probability of this value: p = P ( T & # x2265; 10 ; n = 12 , p = 1 / 2 ) = 0,019 {\displaystyle p=P(T\geq 10;n=12,p=1/2)=0{,}019}
This value is small enough to reject the zero hypothesis and to conclude that the medication works with a blood pressure.
wiki
