HW4

Poli 101
Winter 2020
Homework 4:

Multivariate Regression with Controls (HW4)

Due Date: Pol 101 Exam slot (Monday 16 March 8-11am). Please note that this assignment must be submitted on Canvas no later than 11 am on March 16th.

Important Amendment

Students will have 3 options in completing this assignment.

In submitting your assignment please clearly indicate which of the following options you have selected in completing HW4:
1.    Extend HW3 through Elaboration based on lectures and online labs regarding Regression Models;
2.   Revise HW3 in line with feedback received, indicating all changes in italics;
3.   Indicate that you wish to have your HW3 grade also count for HW4.

Please note that irrespective of which option above is chosen the grade assigned for HW4 will not be lower than that for HW3. Of course, in the absence of a submission for HW4 a grade of zero will be assigned.

Generally speaking, comments will be made only on those assignments attempting hierarchical regression. Otherwise, a grade will be assigned with only minimal written comments. 

Details on Option 1 (Extending HW3 with Elaboration)
Students may work alone or in pairs on this assignment. Student pairs should make only one submission under the name of either partner but clearly indicate both names on the submission.

In extending your work, you are to explore one (or more) aspect(s) of the elaboration model using the hierarchical approach to regression.

Part of this assignment essentially recapitulates what you have already done in HW3. Thus, you will again conduct a simple multiple regression analysis including three (or more) independent variables and a dependent variable which preferably is an index. For this part of the assignment you may repeat much of your work from previous assignments, or you may select new variables.

If you choose to extend your work using the hierarchical approach add at least one (but two is preferable) control variables  to your regression. As always, you should have a clearly formulated theoretical basis for this analysis, stated in this case as a series of hypotheses. Your results should be presented in multi-model regression tables supplemented, where appropriate, with arrow or line diagrams.

An essential aspect of this assignment is to identify a focal relationship and a control variable (CV) from the same data set for inclusion into your analysis. In searching for control variables, please review the elements of the elaboration paradigm discussed in class. A summary is available here (Summary of Statistical Elaboration). Your goal is to find a variable that you reasonably think will explain, interpret or specify the relationship between your dependent variable (DV) and one or more of your independent variables (IVs).

In planning your analysis, it is recommended to focus on a specific relationship (between a DV and a single IV) into which the CV will be introduced. This CV may, but need not, be one of the independent variables used in your previous assignments. In any case, provide the item number and full question wording. Please note that items hypothesized to lead to replication are not appropriate for this assignment.

Your goal should be to describe meaningful results using interpretation, explanation, specification (interaction) or suppression through regression analysis. A clear hypothesis should be the basis for whichever type of elaboration you pursue. Nevertheless, a finding of replication is acceptable, so long as the hypothesis was not one of replication.

In preparing this work, recall that with replication your independent variable will remain a substantial predictor of your dependent variable after controlling for one or more theoretically relevant control variables. In contrast, with explanation or interpretation you will find that by controlling for a theoretically relevant variable (or combination of variables) your original independent variable is no longer as substantial a predictor of your dependent variable as it was originally. In such instances, be sure to consider carefully the logical order of your variables so as to distinguish between explanation and interpretation. Hypotheses regarding specification should be approached via an interaction approach. This entails adding a carefully constructed multiplicative (interaction) term (as well as its constituent parts) into the regression equation.

Irrespective of your theoretical model, the results should be presented in properly formatted, presentation quality tables, including titles and labels. Also be sure to include your syntax.

As in all assignments, do not use examples provided in the lectures, labs or other course material.