Question

5 of 8 patients were recruited to analyze the effects of various varia bles in explaining Total Cholesterol (mg/dL). Some of the variables included in this study are Age in years at screening, Weight (kg), Gender, Triglyceride (mg/d L), Direct HDL-Cholesterol (mg/dL), Fasting Glucose (mg/dL), and Systolic Blood Pressure (1st reading). A multiple regression was conducted to determine the extent to which Total Cholesterol (mg/dL) can be explained by these variables. Here is the output obta ined when a multiple regression analysis was performed, with all of the variables entered simultaneously: Model Summary Adjusted R Square Std. Error of the Estimate Model R Square 346 36.643 1 120 .083 a. Predictors: (Constant, Systolic: Blood pres (1st rdg) mm Hg. Age in years at screening, Trigyceride (mg/d, Direct HDL-Cholesterol (mg/du, Gender Fasting Glucose (mg/dL, Weight (kg ANOVA Sum of Squares df Mean Square Sig .003 Model 1 Regression 30907.959 7 4415.423 3.288 1342.744 Residual 226923.679 169 Total 257831.638 176 a. Dependent Variable: Total Cholesterolk mg/dL b. Predictors: (Constant, Systolic: Blood pres (1st rdg) mm Hg. Age in years at screening. Triglyceride (mg/dL, Direct HDL-Cholesterol (mg/du, Gender, Fasting Clucose (mg/dL), Weight (kg) Coefficients Standardized Coefficients Unstandardized Coefficients Std. Error Beta Model Sig 1 (Constant 189.498 30.740 6.165 000 Age in years at screening 509 152 245 3.355 .001 Weight (kg) 138 117 091 1.180 240 Gender 4.973 5.660 065 879 381 Triglyceride (mg/d -.010 .021 -.035 458 647 380 1.796 Direct HDL-Cholesterol (mg/dL) 211 132 074 Fasting Glucose (mg/dL 229 .145 119 -1.575 117 Systolic: Blood pres (1st rdg) mm Hg -.342 180 .148 -1.898 059 a. Dependent Variable: Total Cholesterolk mg/d

Answer 1

14. The proportion of variance explained in total Cholesterol by independent variables=34.6%

15. Option D: Triglyceride (mg\dL)

16. Option A: Age in years at screening.

17. We use absolute value of beta coefficients or standardized coefficients since the standardized regression coefficient equals the correlation between the independent and dependent variables. Standardized coefficients tell us how increases in the independent variables affect relative position within the group.