Statistical Selection of Effective Body Measurements: Waist Circumference Measurement Exercise

Authors : Nefise Şengöz, Fatih Zeybek

Pages : 50-67

View : 88 | Download : 103

Publication Date : 2023-12-31

Article Type : Research Paper

Abstract :In this research, normality assumptions of body measurements are conducted according to five criteria, boxplots, and Kolmogorov-Smirnov and Shapiro-Wilk tests and proved normal. Besides, statistical selection of effective body measurements to define other body measurements is proposed for shorter, precise, time-cost saving. To support this proposel, correlation coefficients and multivariate multiple regression analyses (MMRAs) are conducted. Statistical selection of effective body measurements depend upon high correlation coefficient pairs of body measurements bases. MMRAs are handled from three different points of view; first, one body measurement being the dependent variable and other body measurements being independent variables enter Method; second, the same stepwise Method; and third, one body measurement being the dependent variable and statistically selected effective body measurements being independent variables enter Method; finally their Adj.r2’s are compared. This research progressed by working on each body measurement, but the waist circumference measurement is only presented here. Enter and stepwise methods concluded the highest and similar results, unexpectedly proposing the least statistical selection of effective body measurements. More profound work with different mathematical sciences will be implemented further.
Keywords : Beden Ölçüsü, Normallik Varsayımları, Korelasyon, Çok Değişkenli Çoklu Regresyon Analizleri, İstatistiksel Seçim