American Community Survey and Census
A lot of data on individual Americans can be found in data from the US Bureau of the Census. The last Census available is from 2000. There are also large annual survey samples, called The American Community Survey (ACS). This includes more than the Census, covering topics such as income, employment, family, education, status, housing costs, and housing conditions. The ACS samples about 1 in every 40 addresses every year, or 250,000 addresses every month.
If you use the ACS (or Census) for this project, you will need to download the datasets with data on each person/household in the survey. You might want to do this on a subset (e.g. only people in a certain state or region, only people with certain kinds of education, only people within a certain age range etc.) depending on your topic, since the entire US data base is huge.
The easiest way to get data is to start at: https://usa.ipums.org/usa/ You have to register for IPUMS (it’s free for students). You want to create your own custom data set. (Do NOT choose to analyze data on line.) Then, I suggest that you watch the tutorial videos (click User’s Guide, then Tutorials on using the IPUMS data and website.) Particularly, watch:
- How to use the Extract System (since you want to extract a data set)
- How to open an extract in Stata
- How to attach characteristics in your extract
- How to use our documentation and variable descriptions
The IPUMS files are a set of individual records. There are two kinds if PUMS ACS files:
1) Files where each observation represents a person.
2) Files where each observation represents a household.
You are most likely going to want to use the person-level ACS files, and to attache variables from the household records.
The files available are:
- Either by state or for the US as a whole.
- Either for 1-year, 3-year, or 5-year periods. The 3-year and 5-year PUMS files are multiyear combinations of the 1-year PUMS file with inflation adjustment factors added.
Click for a list of ACS/Census variables which also lists which years’ ACS/Census has each variable.
GET ANY VARIABLES YOU THINK COULD BE VAGUELY RELEVANT, since it is a pain to find out you forgot to get something. THIS INCLUDES ANY VARIABLES THAT WOULD AFFECT THE VARIABLES YOU ARE MOST INTERESTED IN.
