Homework 4: Computer Exercises

Due Friday Feb 18

Econ 29000

Kevin R Foster, CCNY

For Computer Exercises, your study group should hand in a single assignment. When submitting assignments, please include your name and the assignment number as part of the filename. Please write the names of your study group members at the beginning of your homework.

Remember, I DO NOT want just the SPSS output – although you can include that as an appendix. I want clearly written short answers to these questions, with appropriate charts or tables.

This time we'll use the NHANES dataset, nhanes_forHW4 online.

1. What are the names of the people in your study group?

2. What fraction of the people in the dataset admit to using some kind of illegal drugs (combine the questions for the various possibilities)? More men or women? What about smoking? Alcohol?

Next we will group people by family or household income into quartiles – groups where each is approx. 25% of the total. First use "Analyze \ Descriptive Statistics \ Frequencies" to see that the 25^th, 50^th, and 75^th percentiles are near $20,000, $45,000, and $65,000. [For exact values, click "Statistics" and check "Quartiles."] Create 4 groups based on either one of these measures (family or household income).

3. Now examine the extent to which the conditional means of drug use vary by income quartile. Create a table showing the marginal probabilities. Would you say that illegal drugs are normal or inferior goods?

After drugs, look at sex. The data include details on the number of same-sex and opposite-sex partners for males and females but for ease of comparison these are just added up, as well as top-coding for "30 or more" since about 90% of those who answered this question respond with a number below 30.

4. Plot a histogram ("Graphs \ Chart Builder" then pick "Histogram" from the "Gallery" list at the bottom; drag the first histogram to the canvas, then pick the data item from the list at left) of the number of sex partners. Compare this with summary statistics (mean and median as well as standard deviation) of the number of sex partners without the top-coding. What do you learn from these? How does drug use change with income?

5. [extra] What is the correlation of these decisions on sex, drugs, alcohol, and smoking? Are the correlations different for men or women? For different income or race/ethnicity groups? What about the propensity to answer (particularly for sex partners) – how does this change your answers?