Homework

The data that you will analyze is from a random sample of 300 fictitious cities of varying sizes. The data include information on policing, crime, voting, public services, and structural/demographic characteristics.

Part 1

Using SPSS and the data found in ‘ CRJ535 DATA.sav ’, answer the following questions and complete the following tasks.

1. First, compute homicide rates (per 100k residents) for each city in the data set.
   1. Report the mean, median, and standard deviation for this new variable.
      Next, conduct an independent samples t-test that will determine whether those cities that experienced an increase in population during the past 5 years have mean homicide rates (per 100,000 residents) that are significantly different than cities whose populations have declined during the past 5 years? (Use alpha .05).
   2. Formally state your hypotheses.
   3. What do you conclude on the basis of Levene’s test (Use alpha .05)?
   4. What is the t value and what is the sig. (2-tailed) for the appropriate test?
   5. What do you conclude on the basis of this test?
   6. Report 95% confidence intervals for the difference between the two population means.
2. Conduct an ANOVA that will determine whether the mean voter turnout rate (i.e., the mean of the % of persons who voted in the most recent election) varies significantly by the type of ballots that a city uses for its elections? (Use alpha .05).
   1. Formally state your hypotheses.
   2. What is the appropriate test statistic and sig level?
   3. What do you conclude on the basis of your results (if necessary, consider conducting post-hoc analyses to further enhance your findings).
3. Use Pearson’s correlation to explore the relationship between city-level homicide rates (per 100k residents) and the % of city residents who voted in the most recent election.
   1. Report the value of Pearson’s r and the sig. level.
   2. What do your findings indicate about the nature, strength, and direction of this relationship?

Next, use Pearson’s correlation to explore the relationship between city-level homicide rates (per 100k) and the % of city residents living below poverty. What do your findings indicate about the nature, strength, and direction of this relationship?

1. Report the value of Pearson’s r and the sig. level
2. What do your findings indicate about the nature, strength, and direction of this relationship?

1. You are interested in the extent to which city-level unemployment rates (per 100 persons) influence (or can be used to predict) city-level assault/robbery rates (per 100k residents). Conduct a bivariate regression analysis that will provide this information and present a brief report/discussion of all relevant findings. (Use alpha .05).
2. Use Chi Square analysis to explore the relationship between the region of the U.S. in which a city is located and whether or not its residents can legally carry a concealed firearm.
   If you discover a relationship, how would you characterize its strength?
3. Convert the variable " region " into a dummy variable named "south" that is coded 1 for cities located in the South and coded 0 for cities not found in the South.
   Next, conduct an analysis that will determine whether the mean homicide rate (per 100k) for Southern cities is significantly different than that of non-southern cities. (Use alpha .05).
   Report and discuss all relevant findings.
4. Why is it important that the distribution of residuals from a multiple OLS regression be normally distributed?
5. Run a multiple OLS regression in which the city-level measure of the assault/robbery rate per 100k is the dependent variable. Use the following variables as independent variables: community_policing , zero_tolerance , and education.
   Respond to ONLY the following:
   1. What is the equation for the regression equation?
   2. Interpret the regression coefficient for the variable " zero_tolerance "
   3. Interpret the regression coefficient for the variable " education "
   4. Of the 3 variables in the model, which exerts the greatest influence on the dependent variable? How do you know?
   5. How much of the variability in assault/robbery rates were you able to explain using these 3 independent variables?
   6. Is the amount of explained variation significantly greater than zero? How do you know?

Part 2

For the second part of the homework, you are asked to model city-level homicide rates (per 100k) using multiple OLS regression techniques (again you will use the data from ‘CRJ535 DATA.sav’).

Your task is to develop/build a multiple OLS regression model that meets the following criteria:

1. Explains a large proportion of the variation in city-level homicide rates.
2. Has little or limited problems with multicollinearity.
3. Includes variables that are theoretically related to homicide rates.
4. Includes both interval/ratio & dummy variables.
5. Doesn’t include coefficients for non-significant variables in the model unless their inclusion is justified (you must present a justification).
6. Variables that are relevant to homicide are not excluded without justification (again, you must explain why any variables that would seem to be related to homicide [based on theory or common sense] are not in your final model).
7. You must explore at least one non-linear relationship or interaction effect and report your findings. Depending on your results, these findings may not make their way into your final regression model.

Once you have estimated your final regression model, you will produce a write-up that explains what you did and how you did it (i.e., you will explain/describe the steps that you took to come up with this final model). This includes how and why you created/recoded variables, why a specific category from a nominal variable was selected to be the reference category, as well as your justifications for including, excluding, and removing various variables from your model.

Price: $41.92

Solution: The downloadable solution consists of 23 pages, 1892 words and 31 charts.
Deliverable: Word Document