Comparación de Métodos de Imputación de Datos: Aplicación en el Análisis de Hemoglobina y Factores Asociados ENDES-2023

This study focuses on missing data imputation in the 2023 Demographic and Family Health Survey (ENDES) in Peru, specifically regarding variables related to hemoglobin levels in children under six years old. Different imputation methods were applied and compared using the MICE package in R-Studio, evaluating their impact on the quality and accuracy of the imputed data. Five methods were analyzed: case deletion, mean imputation, regression imputation, random imputation, and multiple imputation. Missing data were mainly distributed across hemoglobin, weight, height, and birth weight. Results showed that mean imputation generated constant values, distorting data variability. Regression and random imputations provided a better approximation of the observed data, with regression imputation showing greater consistency with the original trends. This study highlights the importance of selecting the appropriate imputation method to improve the quality of statistical analyses in health studies.