Regression 2 In this example, we draw on data from the World Bank on four variables for a sample of 47 countries: average life expectancy in years; the population growth rate, as a percent; Gross National Product per capita, in thousands of dollars; and the percentage of people with access to safe drinking water Model 1 Model 2 Model 3 Model 4 Intercept 75.04 68.77 $3.52 $7.81 (1.38) (0.77) (2.18) (2.88) Population Growth -2.65 -0.76 (0.83) (0.60) GNP per capita (sk) 0.35 0.09 (0.06) (0.07) Safe Water 0.24 0.18 (0.03) (0.04) N 37 37 37 37 df 35 35 35 33 R2 0.23 0.48 0.67 0.71 RMSE 4.61 3-77 3.02 2.90 Table 4: Linear regression models of life expectancy. Entries are linear regression coefficients with standard errors in parentheses.Question 6 [1 point} Table 4 collects the results from four separate linear regressions. Focus for the moment on the bivariate Models 1, 2I and 3. According to these models, what is the effect of a 1) two percent increase of Population Growth on Life Expectancy? 2) A ten thousand dollar increase in GNP per capita on Life Expectancy? 3) A ten percent increase in access to safe water on Life Expectancy? Q We cannot say from the above table 0 1) 2.65 year decrease, 2} .35 year increase, 3) .24 year increase 0 1) 5.3 year decrease, 2) 3.5 year increase, 3) 2.4 year increase 0 1) 5.3 year increase, 2) 3.5 year decrease, 3} 2.4 year decrease Question 7 (1 point) Which of the effects identified in previous question are statistically significant at the .05 level? O Population Growth OPopulation Growth , GNP per capita, Safe Water OPopulation Growth , GNP per capita GNP per capita, Safe Water O GNP per capita Safe Water O Population Growth , Safe WaterQuestion 8 [1 point} Which of the following are true of the t-statistic? It is obtained from the normal {Gaussian} distribution It is obtained from the t distribution It is used in hypothesis testing It is used when calculating the standard error of the mean It approaches the z-statistic in large sample sizes In small samples sizer it will be smaller than the z-statistic