A Comparison between Regression and Ratio Estimators using Auxiliary Information: A Case Study of Ladoke Akintola University of Technology, Nigeria

Abstract

This study investigated the use of separate and combined stratified random sampling to estimate population mean scores for two important courses in Statistics using their pre-requisites, by comparing two estimators (ratio and regression). For Probability Distribution, the separate ratio estimator emerged as the optimal choice, providing a mean score estimate of 31.66 with a variance of 6612.18. This indicates that using the pre-requisites course score as auxiliary information improved the estimation accuracy compared to the combined ratio estimator. In contrast, Statistical Inference, the combined ratio estimator proved to be more effective, yielding a mean score estimate of 33.99 with a variance of 84.54. The separate regression estimator was also evaluated, demonstrating its suitability for Probability Distribution with a mean score estimate of 32.18 and a variance of 7.22. However, for Statistical Inference, the combined regression estimator offered a more precise estimate (32.99) with a lower variance (17.59). The findings highlight the effectiveness of auxiliary information when selecting estimation methods in stratified sampling.