Comparative Analysis Of Simple Random Sampling And Multi-Stage Sampling In Estimating National Unemployment Rate

Abstract

This study analyzes the comparative effectiveness between Simple Random Sampling (SRS) and Multi-Stage Sampling methods in estimating the national unemployment rate in Indonesia. Measuring macroeconomic indicators such as the Open Unemployment Rate (TPT) requires high precision but is often constrained by vast geographical coverage and budget limitations. The primary data used is from the National Labor Force Survey (Sakernas) February 2026, which recorded a TPT of 4.68%, a decrease from 4.76% in February 2025. This study also reviews the theoretical foundations of statistical inference, including the use of 95% confidence intervals and the Design Effect (Deff) concept. The results show that although SRS offers an unbiased theoretical basis with minimum variance in homogeneous populations, the Multi-Stage Sampling method applied by the Central Bureau of Statistics (BPS) proves to be far more efficient logistically and cost-wise for a national population spread across 514 districts/cities. The study concludes that the use of staged sampling is able to suppress nonsampling errors through concentrated field supervision without significantly compromising data reliability, making it the standard method for macroeconomic policy.