Political Research Final

What is an appropriate test of significance for relating two nominal level variables?

Chi-square Test

What are the appropriate measures of association for this situation?

Standard measure for association between two categorical variables.

What are the assumptions made in Chi Square?

1. Sample size > 50

2. Each expected cell frequency > 5

3. Independent random sample

4. Nominal/ordinal level data

Properties of Phi

1. Only good for 2x2 table

2. No precise interpretation

3. Results resemble what you would get with Pearson R

Properties of Cramer’s V

1. Good for any size table

2. No precise interpretation

3. The result resembles what you would get with Pearson R

Properties of Yule’s Q

1. Only good for 2x2 Table

2. No precise interpretation

3. Derived from Gamma (ordinal level data)

4. Will give you the highest result of any measure of association we are covering. Note: Phi = 0.18

Properties of Lambda

1. Good for any size table

2. Has a precise interpretation

3. Makes a difference which is in the independent and dependent variable

4. Will give you the lowest result of any measure of association we are covering. Note: Cramer’s V = 0.20

5. Will sometimes give you a false zero

What is research design?

A research design is a strategy or an outline of how a research study is proposed to be conducted to address a specific research question.

What elements does a research design possess?

1. Introduction

   1. Briefly describe what you are about to do and why

2. Review of Literature

   1. Show that you have read everything on the topic

   2. Show the hole in the research and how you will do it

3. Theory, Hypothesis and/or Research Questions

   1. What theories guided your research

   2. What questions or hypotheses do you have

4. Methodology

   1. How are you going to do the research

   2. How are you going to analyze the results

Institutional Review Board

required for any institution that gets federal funds for research, must have an IRB

• Informed consent - must be informed of the benefits and consequences

• Confidentiality

• Safety

• Contribution to society outweighs negative consequences

Research Ethics

1. Concern for the impact of one’s research on society

2. Concern for the impact of one’s research on the subjects

   1. Psychological aspects

3. Institutional Review Board

4. Honesty

   1. Some studies are fraudulent

Examples of Ethical Dilemmas 

Zimbardo Prison Experiment and the Milgram Experiment