ACS 2005-2007


The American Community Surveys (ACS) are random samples of the entire U.S. population conducted every year, beginning in 2000. Separate segments within geographic strata of the U.S. are sampled within the various years. Taken together, the American Community Surveys will cover the entire country in a ten year period, and they are designed to replace the long form of the U.S. Census.

This SDA file of ACS datasets includes the surveys from 2005, 2006, and 2007. It has 8,842,783 person records.

Each ACS has the following characteristics:

  1. It is a national random sample of the population.
  2. The data do not include persons in group quarters.
  3. No place smaller than a state can be identified.
  4. It is not a self-weighting sample. Weights must be used to produce accurate statistics.

All of the ACS questionnaires are very similar, and the datasets from the various years contain mostly the same variables. However, each year of the ACS contains a few questions that are unique to that year.

This ACS dataset is hosted on the SDA server for the IPUMS Project at the University of Minnesota. See: