Designing your own project makes it more enjoyable and more educational. If you have little interest in sports, loss reserving, or interest rates, design a project on social issues, politics, science, business, or any other subject. If you design your own project, it will be accepted, as long as you demonstrate that you understand the statistical techniques that you use.

Don’t fear that you won’t be able to validate your hypotheses. The student project demonstrates that you can apply the statistical techniques to real data. We do not judge whether your conclusion is correct or whether you have used the best explanatory variables. We judge if you can apply the statistical tests to evaluate a hypothesis. Most student projects leave out many important explanatory variables; that is fine.

This candidate regresses youth homicide rates on school drop-out rates, single parents, and poverty. She notes that other potential explanatory variables could be used, but they proved harder to collect.

Social scientists have attributed homicide rates to many explanatory variables, and much controversy continues on each item. In your own student project, you might look at

This candidate compares countrywide youth homicide rates over a period of 21 years. You might design your own student project two other ways:

The ARIMA process fit to the residuals is a proxy for the missing explanatory variables.

The vociferous public debates on these explanatory variables makes it easier to get good data. Your student project shows how much of the social debate relies on dubious assumptions that are not supported by the data. Hundreds of studies have been done to explain crime rates, and you can find good data on many web sites.

Illustration: One social scientist suggested that easier access to abortion reduces the number of unwanted children who lack stable families. These children are more likely to drop out of school and turn to crime and drugs, raising youth homicide rates. Other social scientists say his conclusions are wrong.

Illustration: Some people say easier access to guns raises homicide rates and gun control laws can reduce homicide rates. Other say the opposite is true.

Illustration: Ethnic mix and drug sales are high priority concerns for law officials in California and south-western states. Youth homicide rates rise rapidly if Hispanic and black street gangs occupy adjacent turfs and compete for lucrative drug markets.

Illustration: The effects of policing methods, sentencing laws, and incarceration rates are disputed. Opinions differ about what methods are best to reduce youth homicide rates.

Don’t try to do a perfect study. There are too many potential explanatory variables and too much uncertainty. Examine various web sites and pick data that interest you. Graph relations among different variables and you will soon get ideas for a student project.