In 2001 Kenkel and Terza published an article in which they investigated the impact on an individual?s alcohol consumption of a physician?s advice to reduce drinking. The simplified model below was estimated with their data set (although the econometric techniques used in the article are more complex than those we have discussed in this class).Drinks = 13.00^2 + 11.36 Advie/(5.37) – .20 Educ/(-.65) + 2.85 Divsep/(1.11) + 14.20 Unemp/(2.75)N = 500, R2 = 0.377, t-Statistics in parenthesesDRINKS = drinks consumed by the individual in the last two weeksADVICE = 1 if a physician had advised the individual to cut back on drinking alcohol, 0 otherwiseEDUC = years of schooling of the individualDIVSEP = 1 if the individual is divorced or separated, 0 otherwiseUNEMP = 1 if the individual is unemployed, 0 otherwiseWhich of the following best reflects the interpretation of the coefficient of determination, R2?37.7 percent of the variation in DRINKS is due to random (unexplained) variations.37.7 percent of the variation in DRINKS is explained by variation in the independent variables.37.7 percent of people listen to the advice of a physician in regards to their drinking behaviors.37.7 percent of the people in this sample drink.