Using Stata for Survey Data Analysis

set more off/on is used to turn on and off the continuous scrolling of output. Use “set more off” if 
you are not interested in the intermediate output, only the final result. Use “set more on” if you need 
to be able to read the early output. Remember that the Results Window only stores the most recent 
300-600 lines of output. Unlike SPSS, Stata does not automatically store all of your output.
set scrollbufsize XX is used to change the amount of output that Stata will store in the Results 
window. XX is expressed in bytes. The default is 32,000 (32k) and the maximum is 500,000 (500k).
Type “help set” for a list of other settings in Stata. 
Exercises for exploring the BLSS
Here are some questions that you can answer using the BLSS files provided on your computer and the 
commands described in this section. The file households.dta contains summary variables calculated 
from various other data files. It is at the household level. Open the file by entering “use household” 
in the Command window and pressing “Enter.”
1. How many variables and how many records are in the file households.dta? (Hint: use 
“describe”)
2. What percentage of households have female heads? (Hint: tab b21_q1) 

---

Minot
 
 
 
 
 
Using Stata for Survey Analysis 
Page 16 
 

3. Is there a statistically significant difference between the percentage of female-headed 

households in urban and rural areas? (Hint: use tab command with the chi2 option) 
4. What percentage of urban households are female headed household? (Hint: use “if 
stratum==1” option) 
5. What percentage of female households are in urban areas?
6. How does the percentage of female headed household vary across dzongkhags? 
7. What is the average size of a household? 
8. What is the average size of an urban household in Paro? (Hint: Paro is dzongkhag #3) 
Note: The purpose of these exercises is to practice the Stata commands. To get the correct answers, 
we would have to use the sample weights which are described in Section 7. The weights compensate 
for the fact that some types of households are over-represented in the BLSS sample and others are 
under-represented.

---