Assignment 4

use crime4.dta for this exercise (you can use the file attached, or type bcuse crime4 in stata command section and the dataset will be downloaded from the web). Crime4 has data on 90 counties in North Carolina, for the years 1981 through 1987. Consider the model of crime, where the crimes per person (crmrte) depends on the estimated probability of arrest (prbarr), estimated probability of conviction (prbconv), probability of serving time in prison (prbpris), average sentence length served (avgsen), number of police officers per capita (polpc), and time effects. 

i) get summary statistics of the variables in  the model and comment on them. 

ii) estimate the model using fixed effects and interpret the results (stat.significance, signs, parameters). Do all the variables have the expected signs? Explain.

iii) reestimate the model using random effects and interpret the results. Do all the variables have the expected signs? Explain.

iv) run Hausman test and choose the fixed effects or random effects model based on it.