METHODOLOGY FOR SELECTING PARAMETRIC AND NONPARAMETRIC TESTS IN THE ASSESSMENT OF REGIONAL BUSINESS INDICES

Abstract

This study presents a methodology for selecting parametric and nonparametric tests to assess regional business development using composite indices for Uzbekistan’s 14 regions in 2023–2024. The analysis examines paired observations, normality assumptions, standardisation effects and the economic interpretation of statistical results. The Shapiro–Wilk tests rejected normality for 2023 (W = 0.8595, p = 0.0300), 2024 (W = 0.8378, p = 0.0152) and paired differences (W = 0.8569, p = 0.0276). The paired t-test yielded t = 0 and p = 1.0000 due to annual centring, which does not establish the absence of economic growth. The Wilcoxon signed-rank test (W = 45, p = 0.6698) provided insufficient evidence of a shift in location. The proposed algorithm considers research objectives, sample dependence, measurement scale, distribution, variance, outliers and statistical power. The study emphasises reporting effect sizes, confidence intervals and test assumptions alongside p-values to improve reproducibility and the reliability of statistical interpretations.