Tran Nguyen Gia Huynh K55CLC3 - 1601015181 Average value of prpblck 0.113486 – SD: 0.182416 - proportion of black people in New Jersey and Pennsylvania in percentages Average value of income 47053.78 – SD: 13179.29 - income is measured in dollars Psoda = 0.95632 + 0.114988*prpblck + 1.6E-06*income n= 401, R - SQUARED= 0.06422 The coefficient on prpblckindicates that price of soda increases by 0.115 dollars, or about 11.5 cents, when the proportion of black people in the community increases from zero (no black people) to one (all black people) Psoda = 1.037399 + 0.064927*prpblck n= 401, R - SQUARED= 0.018076 The discrimination effect is estimated to be significantly smaller when income is excluded from the regression (0.064927 vs 0.114988) Log(Psoda) = -0.045678 + 0.111118*prpblck + 1.56E-06*income n= 401, R - SQUARED= 0.066081 If prpblck increases by 0.20, then the price of soda is predicted to increase by 0.111118*0.2 = 2.22% Log(Psoda) = -0.072912 + 0.086163*prpblck + 1.97E-06*income + 0.15052*prppov n= 401 R - SQUARED= 0.070581 The coefficient of prpblck decreases from 0.111118 to 0.086163 Given the results of the above regressions (and intuition), I expected a negative correlation I not have enough intuition to expect a particular magnitude of the correlation (although I was not surprised to see a large correlation between the two) When I found the correlation using Stata’s correlate function, I get: -0.838467 My concise evaluation: this statement is nonsense They can certainly be in the same regression Their simultaneous inclusion, however, makes it more difficult to identify the independent effects of each on the price of soda This does NOT mean that it is illegitimate to include both in a regression