Assignment 5 Solved: Simple regression

Assignment 5: Simple regression

To be submitted electronically

Quantitative methods for conflict studies and human rights, ECH3360

There is a debate about whether guns (and gun laws) deter crime or increase crime. For example, the US National Bureau of Economic Research says:^{Assignment 5: Simple regression}

In theory, the effect of gun ownership on crime is ambiguous. If criminals are deterred from committing crimes when potential victims are more likely to possess a firearm, then more gun ownership may lead to a reduction in criminal activity. If instead guns increase the payoff to criminal activity, or simply increase the likelihood that any particular confrontation will result in a victim’s death, then an increase in gun ownership will tend to increase the crime rate.

In this assignment, we will examine cross-country evidence on the relationship between gun ownership and the firearm-homicide rate. Data on gun ownership is from the Small Arms Survey, an annual estimate of international gun ownership from the Graduate Institute of International and Development Studies in Geneva. The variable gun own is the number of guns owned per 100 people.

Data on gun homicides is from the UN Office on Drugs and Crime.   The variable gun homicides is the number of people killed by firearms per 100,000 people. Income data from the World Bank is also included in the data set (the variable income per capita is per capita income in $US), as is an indicator for whether each country is  an  OECD member (the variable oecd, where 1 indicates an OECD member state; OECD members are wealthy countries committed to democracy and the market economy); Assignment 5: Simple regression

1. Plot the homicide rate against the gun ownership rate. Carefully label and format your plot (and think about which variable belongs on each axis).
2. Conduct a regression of the homicide rate on the gun ownership Briefly explain the results and interpret the meaning of the results for the following: R², coefficient, t-value, and p−value.
3. Are there violations of the assumptions for ordinary least squares? If so, what are they? How might they impact the results? Do the results provide evidence that increases in gun ownership cause increases in the gun homicide rate?
4. Conduct a regression of gun ownership on per capita What do the results suggest?
5. Restrict the sample of countries to only rich countries (with income greater than $40,000 per capita) that are OECD members. Conduct a regression of the gun homicide rate on gun ownership for this sub-sample. What do you find?²
6. Restrict the sample of countries to only poor countries (with income less than than $1,000 per capita) that are not OECD Conduct a regression of the gun homicide rate on gun ownership for this sub-sample. What do you find?
7. Why do you think your results in questions 5 and 6 are different than in question 2?

Assignment 5: Simple regression

Assignment 5: Simple regression