Assignment 3                                      DUE  September 6

 

Again, you will be using the file GSS2002PFP.SAV on the CD-ROM that comes with your book.

 

Part 1.

 

                To see how sampling distributions work you will do the following:

Take the AGE of the respondents. Pretend that your sample of 1500 cases is the entire population and draw 10 samples of 300 cases.

                First check that SPSS will do the sampling properly. Click on Transform --> Random Number Generators. Check the Set Active Generator and choose Mersenne Twister.
Then Set Starting Point by checking that and choose Random.    Click on OK.
 and make sure that the dot is next to the Random Number Seed option. (If it is not put there, you will end up with the same random sample over and over.) Then you can start drawing samples

Step 1. Use Data --> Select Cases. Put the dot in the Random sample of cases circle and click on the Sample button. Choose Exactly and write 300 (cases  from the first ) 1500. Click on Continue and then OK.

Step 2. Then obtain the mean by using Analyze --> Descriptive Statistics --> Descriptives.

Step 3. Then reset the selection to the entire “population.” Use Data --> Select Cases. Put the dot in the circle next to All cases and click on OK.

                Continue with Step 1 through Step 3 until you have 10 means.

 

Then look at the mean for the “population.” What do you see?

(The number of cases may be a bit less or more than 300. Don’t worry about it.)

 

 

BEFORE YOU PROCEED TO PART 2 AND 3 MAKE SURE YOU RESET THE SAMPLE TO THE ENTIRE SAMPLE AND YOU HAVE THE RIGHT NUMBER OF CASES.

 

Part 2.

 

                Is there a relationship between sex (SEX) and attitudes towards gun control  (GUNLAW)?

 

Are women more supportive of gun control then men?

 

Run a crosstabulation for the two variables with the appropriate statistics. (What is the dependent and independent variable? Which percentages will you need? Get only the ones that are appropriate.)

 

Is it statistically significant? Explain your results.

 

 

Part 3.

 

                A researcher is interested in the sexual lives of Americans. Is there a relationship between the number of sex partners and the gender of the person? What is your guess? Do men or women have more sex partners on the average?

 

                To test this, you will have to recode the variable PARTNERS so that it indicates the actual number people answered. The categories 0 to 4 are as they should be, but 5, 6 and 7 are intervals. For the new variable NPARTNER use the midpoint of each category as the actual value (e.g., 5 should be recoded into 7.5 because the midpoint of 5-10 is 7.5). Set -1, 8 and the others to system missing. (Use Transform --> Recode --> Into Different Variables.)

 

                Estimate the mean number of partners and put a 99% confidence interval around your estimate. Use Analyze --> Descriptive Statistics --> Explore.  Just ignore the “Factor List” and "Label Cases by" window.

 

                What did you find?

 

                Are men different on the average than women? Use Analyze --> Compare Means --> Independent Samples T Test to test this. (Your grouping or independent variable is SEX. You will have to click on the Define Groups button and write '1'  for Group 1 and '2'  for Group 2.  )  Assume equal variances. What did you find? Is the difference statistically significant? What is your explanation? (Think hard! This is a slightly tricky problem.)

 

 To answer this question you will  find help on pp.433-436

 

 

 

Hand in your answers and your output.

 

 

Please be sure to present your responses in an organized, clear, grammatically correct, paragraph format (i.e. no sentence fragments, bullets, etc.).  Be sure to proof read, spell check, and make appropriate paragraph breaks when necessary.

 

All homework must be handed in typewritten.  Please use an 11-12 point font, double-space and number each page.