Address Based Sampling, short description

The old industry standard of random digit dialed (RDD) sampling is no longer viable. As people have moved from landlines to cell phones and screen all calls based on caller IDs, response rates have cratered.

A new industry standard, called Address Based Sampling (ABS) has emerged as a viable alternative: it is cost effective and elicits acceptable response rates. Here is a very brief description of Address Based Sampling. One begins by purchasing an RDD phone sample, then paying for address matching (usually a 40-50% match rate), available from the post office (USPS). One then mails a full copy of the survey (with a self-addressed stamped envelope) to respondents. The cover letter includes a link to take the survey online – with a Unique ID for each respondent. One then follows up to non-respondents with a postcard reminder that just provides the online link (again, with the Unique ID), and then follows up with phone calls to non-respondents. Sometimes, one augments this method with a small percentage (say, 10-20%) via a probability based online sampling vendor like YouGov.

Further reading -

  • The definition by AAPOR, the American Association for Public Opinion Research (very long, but thorough: At AAPOR, another copy just in case, here.
  • Michael W. Link, Address-Based Sampling, in the Encyclopedia of Survey Research Methods, 2008, here
  • A nice short definition by MSG, the sampling company, here or here.