Stata Best Practices – Quiz 5

Quiz Question Explanations

Question 1

Since the variable “child_caste_cat” has 5 values and the variable “assessed_ely3” has 2 values, this bysort command executes 5*2 = 10 summarize commands. You could either run the bysort command and count the number of summarize commands, or tabulate child_caste_cat assessed_ely3 and count the number of cells with non-zero observation counts.

Question 2

You can specify the top 25 students within each caste category with a single bysort command:

bys child_caste_cat (child_id_rand): gen top25 = (_n <= _N/4)

This command tells Stata to create a new variable “top25” that assigns a 1 to observations that have observation numbers less than or equal to 1/4 of the total number of observations. Since I specified bysort child_caste_cat, Stata repeats this calculation for each caste category. By including “(child_id_rand)” in the bysort command, Stata will sort by child_id_rand values within each child_caste_cat value before generating the values for the top25 variable. The reason why Stata knows to sort within child_caste_cat instead of using child_id_rand to define categories is because I put child_id_rand into parentheses.
Since OBC students are denoted by the value 3 in the child_caste_cat variable, I can calculate the average baseline score of the top 25% of OBC students with the following command:

sum total_bl if child_caste_cat == 3 & top25 == 1

Question 3

You can use the same approach in this problem as in the last problem:

bys child_caste_cat (child_id_rand): gen bottom25 = (_n > _N*3/4)
sum total_bl if child_caste_cat == 3 & bottom25 == 1