Homework 7
PAM 3100: Homework 7
Question 1
Go to the list of datasets located on the website http://wps.aw.com/aw_stock_ie_3/178/45691/11696965.cw/index.html . Right-click the one named “Guns Data (Stata dataset)”, save it somewhere you can access it and then open it in Stata. This dataset contains panel data from 50 states plus the District of Columbia for the years 1977 through 1999. A detailed description is given in the file “Guns Data Description (PDF)” on the same page. We are interested in the effects of “shall-issue” laws relating to concealed handguns. The variable “shall” is 1 if a state has such a law, 0 otherwise. States that have “shall-issue” laws *shall* (i.e., must) give a concealed gun permit to anyone that wants one (and isn’t a felon, mentally incompetent). States that do not have “shall-issue” laws may issue permits, but generally do so only to certain types of people (i.e., their laws are much stricter). Proponents of “shall-issue” laws argue that they will reduce crime, because criminals will be deterred by the knowledge that some people are carrying guns. Critics argue that “more guns means more crime”. 1. The outcome we will analyze is the log of violence (i.e., you need to create a variable such as ln_vio using a command such as “gen lnvio=ln(vio)”. Why is it convenient to use the log of violence as the outcome variable (rather than the number of violent incidents as recorded by “vio”)? [2 points] 2. Estimate a regression of lnvio on shall for 1977: (a) How many observations are in this regression? [1 point] Is this what you expected? [Hint: think about it in relation to the number of states and the number of time periods in this regression] (2 points) (b) What does the coefficient on “shall” imply is the effect of a shall-issue law on violent incidents in 1977 (assuming we believe these estimates)? 3. Estimate a regression of lnvio on shall for 1977, controlling also for incarcrate, density, avginc, pop, pb1064, pw1064, pm1029. (a) How does this
Question 1
Go to the list of datasets located on the website http://wps.aw.com/aw_stock_ie_3/178/45691/11696965.cw/index.html . Right-click the one named “Guns Data (Stata dataset)”, save it somewhere you can access it and then open it in Stata. This dataset contains panel data from 50 states plus the District of Columbia for the years 1977 through 1999. A detailed description is given in the file “Guns Data Description (PDF)” on the same page. We are interested in the effects of “shall-issue” laws relating to concealed handguns. The variable “shall” is 1 if a state has such a law, 0 otherwise. States that have “shall-issue” laws *shall* (i.e., must) give a concealed gun permit to anyone that wants one (and isn’t a felon, mentally incompetent). States that do not have “shall-issue” laws may issue permits, but generally do so only to certain types of people (i.e., their laws are much stricter). Proponents of “shall-issue” laws argue that they will reduce crime, because criminals will be deterred by the knowledge that some people are carrying guns. Critics argue that “more guns means more crime”. 1. The outcome we will analyze is the log of violence (i.e., you need to create a variable such as ln_vio using a command such as “gen lnvio=ln(vio)”. Why is it convenient to use the log of violence as the outcome variable (rather than the number of violent incidents as recorded by “vio”)? [2 points] 2. Estimate a regression of lnvio on shall for 1977: (a) How many observations are in this regression? [1 point] Is this what you expected? [Hint: think about it in relation to the number of states and the number of time periods in this regression] (2 points) (b) What does the coefficient on “shall” imply is the effect of a shall-issue law on violent incidents in 1977 (assuming we believe these estimates)? 3. Estimate a regression of lnvio on shall for 1977, controlling also for incarcrate, density, avginc, pop, pb1064, pw1064, pm1029. (a) How does this