3. Interpreting statistical software output in multiple regression A study conducted at Baystate Medical Center in Springfield, Massachusetts, identified factors that affect the risk of giving birth to a low-birth-weight baby. Low birth weight is defined as weighing fewer than 2,500 grams (5 pounds, 8 ounces) at birth. Low-birth-weight babies have increased risk of health problems, disability, and death. (Source: Hosmer, D., & Lemeshow, S. (2000). Applied logistic regression (2nd ed.). Hoboken, NJ: Wiley.] Suppose that you conduct a similar study focusing on the age, prepregnancy weight, and weight gain of the mothers as predictors of their babies birth weight among 54 low-birth-weight babies. You use a statistical software package to run a multiple regression predicting birth weight in grams (BIRTHWT) from the mother's age in years (AGE), the mother's prepregnancy weight in pounds (MOMWT), and the mother's weight gain in pounds (GAIN). Use the output that follows to answer the following questions, Correlations BIRTHWT 1.000 AGE MOMWT GAIN BIRTHWT Pearson Correlation -0.3799 -0.1164 0.1925 0.0046 0.4020 Sig. (two-tailed) N 0.1630 54 54 54 54 AGE -0.3799 1.000 -0.0864 Pearson Correlation Sig. (two-tailed) 0.1643 0.2352 54 0.0046 54 0.5345 N 54 54 MOMWT -0.1164 0.1643 1.000 -0.1404 Pearson Correlation Sig (two-tailed) N 0.4020 0.2352 0.3113 54 54 54 54 GAIN 0.1925 -0.0864 -0.1404 Pearson Correlation Sig. (two-tailed) 1.000 0.1630 0.5345 0.3113 N 54 54 54 54 Model R R Square 0.171 Model Summary Adjusted R Square 0.121 Std. Error of the Estimate 323.171 1 0.4140 a. Predictors (constant): AGE, MOMWT, GAIN ANOVA Model df Mean Square F 1 Sum of Squares 1078632.4 5221989.3 Two-Tailed Sig. 3.443 0.02350 Regression Residual 3 359544.1 50 104439.8 Total 6300621.6 53 a. Predictors (constant): AGE, MOMWT, GAIN b. Dependent Variable: BIRTHWT Coefficients Unstandardized Coefficients Standardized Coefficients B Std. Error Beta Model Two-Tailed Sig. 1 2632.1972 324.9535 (Constant) AGE t 8.1002 -27.9660 10.1444 -0.3606 0.0000 0.0081 -2.7568 MOMWT -0.4462 1.6694 -0.0352 -0.2673 0.7904 GAIN 4.9680 4.1377 0.1565 1.2007 0.2355 a. Dependent Variable: BIRTHWT The estimated regression equation is: Y- AGE + MOMWT + GAIN + Following is part of a write-up of your results. Using the information in the table just given, fill in the blanks with the appropriate words or phrases that correctly describe the results of your study. Assume that you have nondirectional null hypotheses and are using a significance level of 0.05. Multiple regression was used to determine whether low-birth-weight babies' birth weights could be predicted from their mothers' ages, prepregnancy weights, and weight gain over the course of the pregnancy. The overall regression was (F(3, 50) = V). Only was a significant predictor of birth weight (t = . ). P- and R2 = Suppose that the coefficients of all of the predictor variables you included in the regression model were significantly different from zero (that is, all three are significant). Which of the following comparisons might best help you think about which variable is the most important predictor? Comparing the magnitudes of the B values (the unstandardized coefficients) Comparing the magnitudes of the standard errors of the coefficients Comparing the significance levels of the coefficients O Comparing the magnitudes of the beta values (the standardized coefficients)