Determination of sample sizes when testing parameters of the populations taken from discrete distributions
Author: Latif OZTURK, Habip KOGAK (Kirikkale University, Kirikkale, Marmara University, Istanbul, Turkey)
Discrete distributions; Databases; Sample size; Simulation.
This study aims to investigate which discrete distribution requires less sample size when estimating parameters. For this reason databases are formed from different discrete distributions with different parameter values using simulation. Samples are taken from the formed databases with Simple Random Sampling (SRS) without replacement and with Sequential Analysis (SA). The parameters are calculated from the databases constructed before and are compared with the values calculated from the sample taken. The minimum differences between parameter value and sample value are determined. In addition, these sample sizes between SRS and SA are compared and which sampling method that requires less sample size when estimating the parameters are determined. Finally which sampling method that would be used is suggested' when the population distribution fit different discrete distributions with different parameter values.