A-B split

Method of random sampling that splits a list of names into two equal groups on an every-other-name basis. One group can then be used as a control group and the other as a test panel. The method is widely used for direct mail promotion testing as well as for most other types of research measuring the response of groups of people to various stimuli, such as advertisements, television shows, and new food products. This type of split eliminates the possibility of sampling errors that might occur from taking whole sections of a list-such as the selection of people who live near each other or of people who share a common ethnic heritage. For example, selecting names that start with "O" would heavily bias the sample toward people of Irish descent. See also nth-name selection; split test.