Stata Software required

Lab C HW

SOCY 3115

Spring 20

Reviewing Labs A and B

Use GSS2018V1.dta

1. There is a variable called ancestrs in the 2018 General Social Survey. This variable reflects the respondent’s answer to this question: Do you believe in supernatural powers of your deceased ancestors?

a. Do a “tab ancestrs” and tab ancestrs, nolabel”. Cut-n-paste command and output here.

b. What level-of-measurement is ancestrs?

c. What percent of the sample believes “no, definitely not” when asked whether ancestors have supernatural powers?

d. What percent of the sample believes that yes (either definitely or probably) ancestors have supernatural powers?

2. Make a new variable called believes that =1 if the person believes that yes (either definitely or probably) ancestors have supernatural powers; and =0 if the person believes that no (either definitely or probably) ancestors do not have supernatural powers.

hint #1: use the generate and recode commands.

hint #2: I did something just like this in Lab C when reviewing Stata from Unit 1

hint #3: you need to use the CODES for each of the values, not the labels. (remember, labels are the words, eg. “yes, definitely,” and CODES are the numbers that represent the words, e.g. “yes, definitely” is coded as a 1, so the number 1 should be used in creating the new variable. You can see the codes by typing, “tab ancestrs, nolabel”

a. Show the commands you used to make the new variable.

b. Run “tab believes” and cut-n-paste the command and output here.

c. Using your new variable, what percent of people believe that at ancestors have supernatural powers? Is this the same as Q1d?

3. The variable called sei10 is an index of the socioeconomic status of the respondent. It’s a composite variable that reflects their education, income, and occupational status.

a. Run “sum sei10” and cut-n-paste command and output here.

b. What is the minimum value of sei10? What is the maximum? What is the mean?

c. Make a new variable called sei75 that =1 if the respondent’s sei is at the 75th percentile or above; and =0 otherwise. I’m not giving you many hints for this. You should be able to find the cutoff for the 75th percentile using the sum command with the detail option. Then, use that value to make the new variable using the generate and recode method. Cut-n-paste the code you used here:

d. Tabulate sei75. Cut-n—paste the command and output here.

New Stuff for Lab C

This week we learned how to manually construct confidence intervals. Now we will let Stata construct the Confidence Interval and we need only to interpret it!

Use GSS2018_v1.dta

1. The variable numpartners is defined as the number of sex partners the respondent has had since age 18. Have Stata construct a 95% Confidence Interval that estimates the mean age of all American adults when they had their first child. Cut-n-paste command and output here:

Interpret the 95% CI like we do when we construct CIs by hand.

2. Using the same variable, construct a 99% Confidence Interval that estates the mean age of all American adults when they had their first child. Cut-n-paste command and output here:

Interpret the 99% CI like we do when we construct CIs by hand.

3. Which of the confidence intervals is wider – the 95% CI or the 99% CI? Why? (use what you know about the tradeoff between precision and confidence in your answer).

4. There is a variable in this file called presexok which =1 if the respondent believes that premarital sex is not wrong at all; and 0 otherwise. Construct a 95% Confidence Interval that estimates the population proportion (i.e. proportion of all US adults) who think premarital sex is not at all wrong. Cut-n-paste command and output here:

a. Interpret the 95% CI like we do when we construct CIs by hand.

b. Are you surprised by the estimated interval? Did you think this proportion would be higher, lower, or about this proportion?

5. Construct a 95% Confidence Interval that estimates the population proportion (i.e. proportion of all US adults) who believe in life after death (probably or definitely). Hint: you will need to create a new variable that =1 if the person believes in life after death and 0 if they don’t. Use afterli to create this new binary variable then run the confidence interval. Cut-n-paste confidence interval here:

a. Interpret the 95% CI like we do when we construct CIs by hand.

b. Are you surprised by the estimated interval? Did you think this proportion would be higher, lower, or about this proportion and why?