Determine Sample Size for Precision Results on Quick Count

Abstract

This research aims to answer the problem of the appropriate sample size in the case of the quick count of the election so that the results obtained are close to the actual results. Although there are practical procedures that are widely used to calculate the sample size in the quick count methodology, in reality, the results obtained often deviate from the actual results, so the issue of precision is always an interesting discussion. The formulation of the problem regarding the size of the sample and how the level of precision of the forecast results are important issue to be discussed. This research method is included in experimental research where the analysis used is the Kruskal-Wallis test. The data used is primary data from the real count results of the regency election Sumedang by consultants and teams. The results showed that there was a significant difference between the seven sample size groups in vote acquisition and the percentage of votes at the polling station (TPS), where the sample sizes n=408, n=500, n=875 and n=1674 were the most appropriate sample sizes in the implementation of the quick count.