POLS210

Assignment 14

Regression

[Note: The following instructions are written to help you complete the assignment. (However, they are still not enough to replace the entire lecture. If you missed the lecture, you would probablly have many problems understanding the instructions. If you face problems, try to complete the assignment as much as you can before you come to class and ask the instructors some specific questions.) Reading the required reading may also help you complete the assignment.]

Do not include any wrong or unnecessary output and make sure that the tables are organized well.

The following link shows an example of SPSS output.

1. Use the STATES.SAV dataset (Do not use the student version) and analyze the impact of OVER64 on PROCHOICE and find out if the results support the directional hypothesis that the states with more older people have larger percentages of people who are pro-choice.

(SPSS: Analyze--Regression--Linear ).

Make sure to include the following

1. State the null hypothesis

2. Analyze the impact of Age.

a. Highlight the number that shows statistical significance.

b. Do you reject the null hypothesis?

c. Is the relationship statistically significant?

d. What is the confidence level that this variable is reaching?

e. Underline the number that shows the slope.

f. Is the relationship positive or negative?

g. Read the slope coefficient substantively. Make sure to use the actual number of the slope coefficient in your statement.

h. Do the results support your directional hypothesis?

i. Interpret the adjusted R^{2}. ( R sqd and Adj R sqd will be similar, when the number of cases is large.)

^{ (e.g.
"The model explains _% of the variation in the dependent variable.")}

2. Use the WORLD.SAV dataset and analyze the impact of FRAC_ETH on FREE_CORRUPT after controlling for SPENDEDUC and GDP_10_THOU. Suppose you are hypothesizing that ethnic fractionalization is associated with more corruption. (SPSS: All three X variables should be placed in the "Independent" box.)

1. State the null hypothesis

2. Analyze the impact of Age.

a. Highlight the number that shows statistical significance.

b. Do you reject the null hypothesis?

c. Is the relationship statistically significant?

d. What is the confidence level that this variable is reaching?

e. Underline the number that shows the slope.

f. Is the relationship positive or negative?

g. Read the slope coefficient substantively. Make sure to use the actual number of the slope coefficient in your statement.

h. Do the results support your directional hypothesis after controlling for the education and economic variables?

* Hint: What does the slope coefficient suggest? Think about how the corruption variable is coded. Larger numbers mean less corruption.

e.g.

UK (less corrupt)=77

Russia (more corrupt)=21

Does the slope coefficient suggest a positive (or negative) slope? Draw a regression line below and think carefully.

i. Interpret the adjusted R^{2}