Fill This Form To Receive Instant Help
Fill This Form To Receive Instant Help
Homework answers / question archive / Using census data for the 50 U
Using census data for the 50 U.S. states the following model was estimated (values in parentheses are the standard errors). PC = -0.365 -0.0009D +0.0094NW+0.0003 Y -0.099U + 1.519COP -0.0068AGEI +0.0077AGE2 (0.978) (0.0006) (0.0104) (0.0001) (0.084) (0.276) (0.00034) (0.0038) R-0.557 R? -0.484 ESS - 33.41 where PC - property crime index D-population density NW = percent of non-white population. Y-per-capita income in dollars U - unemployment rate COP = size of police force per thousand population AGEI = Population (in thousands) in the age group 15-24 AGE2 - Population (in thousands) in the age group 25-34 a) What signs would you expect a priori and why? Does any results surprise you? b) Test each coefficient for statistical significance at 5% level. A second model was estimated and the following results are given; (R) PC = -0.37 -0.0002Y+1.428COP -0.0068AGEI +0.0077AGE2 (0.84) (0.00008) (0.266) (0.839) (0.839) R=0.512 R? -0.468 ESS = 36.85 c) Using both the models (U) and (R), perform an appropriate test at %5 level. Be sure to state (1) the null and alternative hypotheses, (2) the statistical distribution (including the degree of freedom), and (3) the test creation for rejection of the null. What do you conclude from your test? d) For model (R), the F-statistic for overall goodness of fit is 11.8 and the corresponding p-value is 0.003. State the null hypotheses for this F-test. From the p-value would you accept or reject the null? Why? e) In Model (R), do the signs of coefficients agree with your intuition? Based on that would you say that the models here are sensible?
e) The signs of coefficients are not always aligning with the intuition.
In the model we see that per capita income has a negative coefficient which makes sense as lower the income higher the chances of property crime. But the value of the coefficient is low.
However, size of police force has a positive relation with property crime. One would believe a higher police force will reduce property crime intuitively.
Next we see that the ages of 15-24 show a negative relationship to property crime while ages of 25-34 show a positive relationship. This is also arguable as we note that ihstoric data shows that there is a high property crime rate in the age of 15-24. In fact the peak age for property crime arrests in the United States is 16, compared to 18 for violent crime arrests.
Thus overall the model does not seem sensible intuitively and there seems to some sort of model mis-specification which has led to counter-intuitive results and poor coefficient estimates.
