Assignment #6: Correlation & Regression

1a. Using the State Court Processing Statistics database, perform the following bivariate correlation: total number of arrest charges (totchgs2) and number of months sentenced to jail (jailmths). What is the direction and magnitude of the association? Is it a significant association? If so, at what level?

1b. Using the same data and the same variables, perform a regression analysis to see if the total number of arrest charges predicts (not just is associated with) length of the sentence. Interpret the Y-intercept (constant in SPSS) and the regression coefficient (Unstandardized B in SPSS). Does the total number of arrest charges significantly predict sentence length? How much of the variance is explained by this one variable? Please give a substantive explanation of your findings.

1c. Using the State Court Processing Statistics database, perform a multiple regression analysis to see how the total number of arrest charges predicts length of sentence after including two more independent variables: age and number of prior convictions (priconv). Interpret the Y-intercept and the regression coefficients (do each predictor separately, regardless of significance level), and using the Standardized Beta coefficients, indicate which predictor has the strongest effect on sentence length. How much of the variance is explained after including age and prior convictions to the model? What is your substantive interpretation now?


2a. Using the College Alcohol Survey, perform the following bivariate correlation: age (a1) and number of drinks consumed in last 30 days (c10). What is the direction and magnitude of the association? Is it a significant association? If so, at what level?

2b. Using the same data and the same variables, perform a regression analysis to see if age predicts (not just is associated with) amount of alcohol consumed. Interpret the Y-intercept and regression coefficient. Does age significantly predict alcohol consumption? How much of the variance is explained by this one variable? Please give a substantive explanation of your findings.

2c. Perform a multiple regression analysis to see how age predicts alcohol consumption after including two more independent variables: gender and Greek membership [1] . Interpret the Y-intercept and the regression coefficients (do each predictor separately), and using the Standardized Beta coefficients, indicate which predictor has the strongest effect on alcohol consumption. How much of the variance is explained after including gender and Greek membership? What is your substantive interpretation now?

[1] Getting tricky here, but think logically and you’ll get it. These variables are dichotomous, nominal-level predictors, not continuous, interval-level predictors as we have been using. Interpret the Y-intercept and regression coefficients the same way (i.e., when X = 0 (female or non-greek Member), Y = …; for a one unit increase on X (going from female to male or non-member to member), a ____ unit increase on Y occurs (how much more males drink than females, how much more members drink than non-members)).

Price: $11.84

Solution: The downloadable solution consists of 5 pages, 684 words and 8 charts.
Deliverable: Word Document