Tutorial: using census microdata to find dates

variable view in SDA

I’ve talked about using the US census to estimate your dating options in Americain Maine, and in selected US cities. If you’d like to explore your own area, here’s how to get started:

1) Set up your account and use the online analysis tool

First, you’ll make a free account with the SDA online analysis tool. I select from a sample of American households in 2010-2014 to get a rough estimate of the characteristics of people living nearby:

SDA American Community Survey getting in

2) Look up your PUMA

The smallest group of people you can look at is a set of 100,000 people in a given area (public use microdata area, or PUMA). In a city, that could be a few blocks. In a rural area, it could be half a state. To find your PUMA, view the US census maps for your state:

finding your PUMA in census geography reference maps

3) Play with the online SDA tool

Once you’ve got the number of your U.S. state and PUMA, go to the 2010-2014 census estimate of to play with the free “SDA” online analysis tool.  On your left is a menu of all the variables the census collects on people. Start clicking those books of data, or enter a name (“marst”) in the Selected box if you know what you’re looking for:

variables marst in US census

If you select a variable and click “view” (above), a tab will pop up with more information on how the census sorts people. So for marst, value 1 = married, value 6 = never married, etc. Notice that they didn’t actually survey all Americans. Over five years, the government asked 15,552,144 people whether they were married, and then used that sample of Americans to estimate how many of us are married, divorced, or single:

variable view in SDA

Knowing these details, I can filter all the people in my area with marst(6) to see only the never-married people. Over time, I’ll add more categories and get more specific. Below, I ask the census to compare men and women (in “column”) by their education level (my “row”). I’ve combined some of the government’s categories (so educ(7-9) all become “some college”), and I’ve also given them handy labels (so sex(2) becomes “gal”):

sample input in SDA for American Community Survey

I’m also using filters to narrow my options: just people in the state of Maine (state 23), in the central area (PUMA 600), who have never married (6), and are age 30-35.

And here are my results. There are twice as many single women as men here. The available ladies are more likely to have some college, and the guys more likely to have finished a bachelors or four year equivalent:

ACS results for single Mainers with some college

Note that I didn’t include a lot of people: married people, those who left high school early, or 36-year-olds. You can set these boundaries however broad or narrow you want, as you study your area. And don’t forget to run this again for another area–comparison is part of the fun!

Bonus! You can also use this SDA interface to explore other free data, such as opinion polls of Americans from 1972 to the present. For more on that, check out our slides below:


Leave a comment

Your email address will not be published.