5.5 Stratified random sampling and judgment sampling

Parent Previous Next


The main difference between stratified random sampling (judgment sampling) and random sampling is that with the former the sample frames are constructed around a specific group of people. First, the researcher must stratify the sample according to the categories which are relevant to the project. Then, he/she must identify a respective group of potential participants of which to randomly choose the informants. This kind of sampling is used in most linguistic research (cf. Milroy and Gordon 2001: 26, 30; Tagliamonte 2006: 23-24).

One problem is that defining the sampling universe is not as straightforward as it sometimes seems. For example, Milroy and Gordon (cf. 2001: 27) point out that if a fieldworker is interested in studying a language in a particular city, he/she should avoid interviewing non-native residents as they may have a distorting effect on the obtained data. On the other hand, these people may have a large impact on language change within this area, so obviously, depending on the research aim, they may be essential to the project which makes defining the relevant categories difficult.

The other problem discussed by Milroy and Gordon is that searching out a defined group may turn out to be more complicated than expected. For example, finding members of a particular minority in a town may prove to be a difficult task. One example of this fact is provided by the London based Linguistic Minorities Project, with researchers affiliated to the project having encountered unforeseen obstacles when trying to locate Italian immigrants. Minorities are spread across society in a non-random way and therefore random sampling is an inappropriate way of addressing them. As a result, two different approaches were taken by members of the Linguistic Minority Project. They tried finding suitable participants by means of an ethnic name analysis and with the help of community lists of minority language speakers. Both approaches proved unsatisfactory, as hardly any Italian names appeared on electoral registers and the ones that did often proved to be unreliable indicators of ethnicity.

More information on how many people to choose is found here.

Created with the Personal Edition of HelpNDoc: Easily create CHM Help documents