Nonprobability sampling
Random sampling ( random selection ) are special selection of empiricism. Elements from the population are included ( approximately by an interviewer ) more or less at random in the sample, it is entirely at the discretion of the interviewer or the panelists themselves, the probability of entering a particular element in the sample, can not be specified. The random sample is therefore not suitable statistical inference ( generalized underestimated ) derive statements. It is impossible to give up, to which population the respective sample relates ( representativeness ).
Problems are distortions in the direction of the selection mechanism and self-selection.
Examples:
- Survey of people in a pedestrian area (English " Mall Interception Technique" ): overrepresentation of participants which speaks to the interviewer and staying in for questioning at the time Befragungsort.
- Open online surveys ( online surveys, where participation is not limited to certain people): overrepresentation of participants who use the Internet and control the survey page.