Week 12 dummy coding

# Week 12 dummy coding - SAN JOS STATE UNIVERSITY College of...

This preview shows pages 1–3. Sign up to view the full content.

SAN JOSÉ STATE UNIVERSITY College of Social Work S. W. 242 Spring 2009 Edward Cohen Weeks 12 April 17, 2009 Review for test Dummy variables in multivariate regression Lab 4: Multiple Regression Small group work on Analysis Plan Writing Abstracts I. Review for test A. Choosing Statistical Tests 1. Know when to use statistical procedures--in the abstract, e.g. knowing level of measurement and number of IVs, DVs; and for specific research scenarios 2. Test will cover these: a. Dependent (or Paired) groups t test b. Independent groups t test c. Multivariate regression (but not logistic regression) B. Hypothesis Testing--as in the first exam, you will be asked to go through the 8 steps of hypothesis testing for a given scenario 1. What is (are) IV(s)? Type of measure 2. What is DV? Type of measure 3. Identify null hypothesis 4. Identify alternate hypotheses 5. Choose statistical test and alpha 6. Interpret SPSS output 7. Justify decision to reject or not reject null hypothesis, and other Results y Know how to interpret both significant and non-significant findings y For t tests: interpret differences in means to support research hypothesis y Find significant regression coefficients; support research hypotheses? 1

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
y In regression, identify IV with largest impact using the Standardized Beta coefficients y Understand how to interpret coefficients for dummy variables 8. Interpret results a. Did findings answer research question(s)? b. Study limitations Research design Sampling strategy and representativeness c. Implications of findings for social work practice and/or policy d. Suggestions for further research II. Dummy Variables in Multivariate Linear and Logistic Regression Dummy variables are transformed nominal or ordinal variables whose attributes are coded into dichotomous variables. A dummy variable is dichotomous, e.g. the variable named “Latino” has only two attributes: 1=Latino; 0=Not Latino. If an independent/control variable is categorical (either nominal or ordinal), then dummy coding is necessary for proper analysis with multiple regression. This involves creating a separate variable for each category (attribute) of the categorical variable and using a “baseline” category with which to compare all other categories. Why do we need dummy variables? Consider “ethnicity”—if coded 1=White, 2=AA, 3=Latino,
This is the end of the preview. Sign up to access the rest of the document.

## This note was uploaded on 09/08/2010 for the course SCWK 298 at San Jose State.

### Page1 / 6

Week 12 dummy coding - SAN JOS STATE UNIVERSITY College of...

This preview shows document pages 1 - 3. Sign up to view the full document.

View Full Document
Ask a homework question - tutors are online