Where is our Copernicus?
And Twitter shall save the world

What is Address Based Sampling?

In the previous post I talked about Address Based Sampling.  It caused a couple of people to write and ask what I was talking about.  There was a similar discussion on AAPORNet which is a closed list, and so I could not simply cross post.  Below is an excerpt from the AAPORNet thread.  It was written by Michael Link from Nielsen who has done a good deal of work in this area.

Address Based Sampling is the sampling of addresses from a near universal listing of residential mail delivery locations. This distinguishes it from more manual counting and listing (or in-person enumeration) efforts typically used with area probability sampling. In the US this is typically done using the USPS Computerized Delivery Sequence File, but the definition is generic enough that it could fit databases in other areas or countries which provide a near universal frame of addresses.

The concept of using the DSF as a potential sample frame for assisting with (or ultimately replacing) counting and listing was first examined by Iannacchione, Staab, and Redden (2003). The first use of the actual term "Address Based Sampling" as it has been used for the past few years in the US when referring to drawing samples of the general population in the US was by Link, Battaglia, Frankel, Osborn, and Mokdad (2008).

Admittedly the general concept of "address based sampling" (or sampling of addresses) has been around for a long time to refer to any sample drawn from a frame of addresses. More recent usage has the more specific meaning.

Thanks to Michael.