Applicability of Regression Equations for Evaluation of the Effectiveness of Physical Education among Students of Higher Education Establishments

Abstract

Methods: Theoretical analysis and generalization; anthropometry; method of least squares; correlation and regression analysis. Materials: We have studied the applicability of the regression equations to determine the effectiveness of physical education using the results of morphological, functional and physical readiness among students of 1st – 3rd years of studies (N = 142) at the main educational departments of technical courses. We have also determined the possibility of evaluating their achievements using the arithmetic mean in eight test exercises. Results: The applicability of the regression equations for evaluation of the results of test exercises has been determined with the coefficient of determination and the percent average approximation error. It has been found that the choice of the regression equation, its digital format, and the way of determining control points influences the assessment accuracy. It has also been found that in order to calculate arithmetic mean values, it is essential that the test exercises create a heterogeneous complex. The heterogeneity of the complex has been determined as a modulus of the values of the partial correlation coefficients which have been used as bases to judge on the multicollinearity of test exercises. Conclusions: The study shows that the assessment of achievements using the test exercises is done with the well-chosen regression equations (D = 99.82 ÷ 100%) with high accuracy of scoring both of every exercise (Ā = 0 ÷ 1.52%) and of the test as a whole (Ā = 0.71%). The calculated partial correlation coefficients have indicated that this test is heterogeneous since the highest value among all obtained values r x5 x7 / x3 = 0.680 is less than the critical multicollinearity
indicator – 0.7.