Comparison of the strength of Modified Extended Median Test for c matched samples and Extended Sign Test by Rankusing a random sample of a certain specie of Bat that were offered dose levels of certain drugs to reduce heartbeat.

Background

Non-parametric tests are those tests normally applied in situations where there are violations of the normality and homogeneity of variance assumptions. In other words,parametric tests cannot be used in such situations. Here, we shall be comparing the statistical power of two non-traditional non-parametric statistical methodssuitable for analyzing data that whenhandled as a two-way Analysis of Variance would be same as a mixed-effects model without replication. They are Extended Sign Test by Ranks for ordered repeated measure and Modified Extended Median Test for c matched samples.

Methods

This paper adopted methods proposed by some authors for Extended Sign Test by Ranks for ordered repeated measure and Modified Extended Median Test for c matched samples. These methods were illustrated using a random sample of a certain specie of Bat offered dose levels of certain drugs to reduce their heartbeat.

Results

On using the method of Extended Sign Test by Ranks for ordered repeated measure to analyze the data, we rejected the null hypothesis of equal dose levels at 5% significant level since

But at 1% significant level, we accepted the null hypothesis, since On using Modified Extended Median Test for c Matched samples on data, werejected the null hypothesis at 5% significant level since  and same also at 1% significant level since

Conclusions

Extended Sign Test by Ranks for ordered repeated measure has much greater chances of committing Type II Error at 1% significant level, hence Modified Extended Median Test for c Matched samples is more powerful than Extended Sign Test by Ranks for ordered repeated measure and is therefore recommended for use.