Using Stata for Survey Data Analysis

freq
label Y axis in terms of frequency 
percent
label Y axis in terms of percent 
normal
draws a normal curve with the means and standard deviation of the variable 
Some options for two-way scatterplots: 
connect( ) 
to specify how points are connected
msymbol( ) 
to specify what the marker look like 
Some options for bar charts: 
over(var1)
plot y-variable for each value of var1 in one graph 
stack
stack the bars for each variable rather than putting them side by side 
Here are some examples of the histogram, scatter, and graph commands:
histogram x
histogram of x 
histogram x, bin(5) 
histogram with 5 bars 
scatter y1 y2 x
scatter plot of y1 and y2 against x
scatter y x, by(region)
scatter plots of y against x for each region
graph bar a b c
bar graph of the means of a, b, and c 
graph bar (sum) a b c
bar graph of the sums of a, b, and c s 
 

---

Using Stata for Survey Data Analysis 
 
Minot 
 
Page 47
We can give an example of the graph command by calculating the proportion of expenditure spent on 
food. After merging the “households.dta” file and the “food expenditure.dta” file, we use the 
following commands: 
gen foodexp = exp_annu 
gen totalexp = pc_t_mo*12*hh_size 
gen foodshare = foodexp/totalexp 
drop if foodshare > 1 
gen lnpce = ln(pc_t_mo) 
scatter foodshare lnpce, by(stratum) 
 
Example 20 shows a scatter plot of the share of expenditure spent on food as a function of per capita 
expenditure. The graph shows that as income (or per capita expenditure) rises, the share spent on 
food declines. This relationship is called Engle‟s Law, and it is one of the most universal patterns 
found in economics.