ANALYSIS OF THE AUTODETERMINATION COEFFICIENT STATISTICAL SIGNIFICANCE

Abstract

The study examines the issue of estimating the statistical significance of the autodetermination coefficient. The features of the proposed estimate are identified – bias, efficiency, consistency and sufficiency, and their relationship with the length of the time series and the number of the autocorrelation coefficients used. Two approaches are proposed for testing the statistical significance hypothesis. The first one is used in long time series and involves calculating the properties of the Wald test (W), the likelihood ratio test (LR) and the Lagrange multiplier test (LM). The second approach is used in short time series on the basis of tabulated boundary values of the autodetermination coefficient. The study proposes an analysis scheme which was used in studying the time series of the main indicators of Bulgaria’s demographic development for the period 1930 – 2011.