Abstract:

The objective of this research is the performance of the methods Mantel-Haenszel (MH), SIBTEST and Substance Reaction Theory-Powerability Rate (MTK-OO) with the simulation work under certain conditions (I. The type of error and power is compared. The variables that are studied on DMF are the sample size ratio, skill in capacity distribution, test length, the substance ratio containing DMF and the model type. The findings from the research show that the highest error in the determination of single-form DMF is the SIBTEST method, and the MH method has the highest power in all the conditions addressed. In the study, however, the DMF substance ratio and model type variables are MTK-OO; capacity distribution, test length, DMF substance ratio, model type variables and capacity distribution*DMF substance ratio, capacity distribution*test length, test length*DMF substance ratio, test length*model type interactions SIBTEST; sample size ratio, test length, DMF substance ratio variables and capacity distribution*DMF substance ratio, capacity distribution*model type, capacity distribution*DMF substance ratio*MHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHH It is found that the type of error has a significant impact on performance. The effect of any condition on the power of the SIBTEST and MH methods is not significant, while the model type variable has a significant effect on the power of the MTK-OO method.

Keywords: