See attached assignment file & GSS data set
SOC3040 Assignment 4 Assignment 4 Social Research: Analysis (SOC 353) Overview This assignment will focus on association, intermediate hypothesis testing, and regression. The breakdown of points is as follows: Topic Number of Points Measures of Association 1 Working with SPSS 9 Total 10 Since the first question requires hand calculation, you do not need to type all of your answers using word processing software. For the question requiring pencil-and-paper calculations, you may submit your handwritten results (write neatly, please). Be sure to show all of your work, since partial credit is possible. For the SPSS component of this assignment, please complete all work using word processing software (e.g., Microsoft Word, Libre Office, etc.). This assignment is worth 10 points, or 10% of your final grade, though a score of 11% is possible if the bonus question is completed. Measures of Association 1. You have been given data to determine the strength of the association between two variables: x y 0.34 0.89 0.64 -0.69 1.51 2.53 -2.00 -0.2 0.48 0.37 0.63 1.75 -1.63 -2.87 0.44 -0.71 0.18 -0.31 -1.02 -1.26 Calculate the appropriate measure of association between x and y, given their level of measurement. Be sure to interpret the strength of the association in light of the 1 thresholds we discussed in class. You only need to keep four decimal places (i.e., 0.0001) in your calculations (1 point) Working with SPSS The bulk of this assignment involves using SPSS for correlation, ANOVA, and regression. Before turning to the questions, keep the following in mind: • All analysis must be completed using the GSS data set. Failure to use the GSS data set will result in a zero for the question. • For the regression question, you must present a professional quality table. Sim- ply copying and pasting SPSS output will result in a 10 point penalty. • Use SPSS for the scatterplot in Question 1 and Excel if you are attempting the optional bonus question. • You must describe any recoding with enough detail to allow replication. • Though you may work with variables of your choosing, you may not replicate any example that we covered in class or provided by Healey. • Create dummy variables correctly using the “recode” procedure we have covered in class. • Be sure to both mechanically and substantively interpret your results. 1. Find any two interval-ratio variables that you believe should be associated. Provide a brief justification for why you believe an association exists and the direction you expect the association to have. Estimate Pearson’s R and produce a scatterplot of the data. Interpret your results, paying attention to both R and the scatterplot. (2 points) 2. Find an interval-ratio dependent variable and a categorical variable with at least 3 categories. Briefly state whether you believe that there are significant differences across the different groups and why. State the null and research hypotheses for your variables and then estimate an ANOVA. Interpret your results in light of your expectations. (1.5 points) 3. Pick an interval-ratio variable of interest that will serve as your dependent variable, and at least three independent variables (at least one of these must be interval-ratio, and one must be dummy variable coded). For each variable that you have selected, briefly state the expected direction (i.e., positive or negative) of the effect and why you believe this to be the case. Estimate a linear regression model, table your results appropriately, and substantively interpret the model (i.e., describe what the results tell you, given your expectations). Make sure that your regression model significantly improves your ability to predict y and that you discuss the % of variance explained by your model, the intercept, each partial slope, and whether the intercept and slopes are statistically significant. You should also make use of at least two predicted values when interpreting your results. (5.5 points) 2 Optional Bonus Question: 1 Point Create an effects display plot in Excel demonstrating the difference in intercepts between two (or more, as appropriate) categories of a categorical variable across the span of an interval-ratio independent variable based on the model you estimated in the previous ques- tion. Interpret this as well. An example of the sort of plot you will need to produce is provided on this page. Make sure to appropriately title the axes of your graph and give it a proper title consistent with the standards of professional statistical graphics. 20 30 40 50 60 70 80 90 30 35 40 45 50 Weekly Hours Worked by Sex Age A ve ra ge W ee kl y H ou rs W or ke d Men Women 3