survey income data by wealth quintile

[Robert Kolesar]

Hello,

I am using Stata 15 to analyze survey income data by wealth quintile for households. Two questions:

1. Do I use the sampling weight with xtile? This seems right, but it seems odd because I am also weighing the pie graph.

2. What svy command can I use to just show me the proportion, total value, mean, and range by wealth quintile?

Here is my code so far.

Thanks much. Robert

xtile quintile_totalincome = hhtotalincome [pweight = HHweight], n(5)
graph pie totalincome_positive_values_US [pweight = HHweight] if HHhead==1, over(quintile_totalincome)
svyset psu [pweight= HHweight]

[Steve Samuels]

You can get weighted quintiles with pctile, as it takes pweights. Once you have them:

svy:  tab quintile_totalincome  // proportions
svy: mean hhtotalincome, over(quintile_totalincome)  // means
svy: total hhtotalincome, over(quintile_totalincome)  //  totals
bys quintile_totalincome: sum hhtotalincome   // min and max

The pie chart show weighted percentages over categories. However with quintile categories, those percentages are equal (20%). Therefore the graph really conveys no extra information that I can see.

[Robert Kolesar]

Thanks a ton! Much appreciated.