Evaluation of panel data estimators under the unbalanced panel data for small data sizes occasioned by missingness
Author(s): OP Balogun, WB Yahya and AA Issa
Abstract: This paper examines the performances of the developed Between Median estimator (BMd) for balanced panel data under the Unbalanced panel data for small datasets by introducing varying degrees of missingness. This study focuses on 5% increments of missingness (5%- 20%), and sample sizes, N (20 and 30) with a unit beta (slope); β=. Adopting the Monte -Carlo investigation of (Balogun et al 2022), where the behaviors of the panel data estimators’ using Mean Square Error (MSE) and Mean Absolute Error (MAE) criteria to measure their performance were tested; among the five estimators of balance panel data models under Unbalanced panel data estimators to varying levels of degree of missingness injected for panel data (N=25) shows that the Between estimator performed best among the estimators. In a separate study, (Balogun and Yahya 2024) [3] found that the BMd estimator, developed for small size N (20 and 25) with β=3 and T = 5 panel data under the Unbalanced panel data has the lowest MSE and MAE values, and therefore, outperformed the Between estimator. This work builds on (Balogun and Yahya 2024) [3]; the findings demonstrate that the developed estimator (BMd) performed best of the six estimators tested using the same measurement criteria.
OP Balogun, WB Yahya, AA Issa. Evaluation of panel data estimators under the unbalanced panel data for small data sizes occasioned by missingness. Int J Stat Appl Math 2024;9(6):144-149. DOI: 10.22271/maths.2024.v9.i6b.1914
