Re: Snowball or Purposive Sampling???

Posted by Mark Schneider on February 20, 2000 at 15:10:19:

In Reply to: Re: Snowball or Purposive Sampling??? posted by suluan on February 20, 2000 at 03:27:44:

Snowball sampling is a form of convenience sampling that allows the researcher to locate very small groups of respondents via identification of them by their friends, relatives or acquaintances who either know them or know of them. It may be used in sampling drug users, people afflicted by rare diseases, etc. It's sometimes even used in recruiting for focus groups of consumers of products having minute market share. And I believe even the Census Bureau may have used a form of snowball sampling during its one-night census of the homeless population prior to the 1990 Census Day in order to locate places where the homeless lived or congregated in order to enumerate them.

A snowball sample may start with a small number of known group members (such as patients at a drug re-hab center) or even a random sample to identify members of the rare group. But it then "snowballs" as those who are interviewed are then asked to provide names of others they know in the group, and those respondents are asked for other names, etc., etc.

Of course, such a sample is loaded with biases and usually not projectable to any larger group beyond the sample itself. But it may be the only affordable or timely way to locate members of the rarified population, or to identify people who may be reluctant to identify themselves to an interviewer, like a heroin user.

Follow Ups:

Subject: