Sweden key

from Wikipedia, the free encyclopedia

The Swedish key (also known as the Kish Selection Grid) is a method developed by Leslie Kish for the random selection of respondents in households with several people. Its use is particularly necessary for surveys with a sample procedure if the addresses selected on site for a sample initially only represent households as a whole ( household sample ). In the next step, every person in the household should be given the same chance to ultimately act as a survey partner by using the Sweden key.

Simple implementation example for a written survey (PAPI - Paper and Pencil Interviewing)

  • A maximum household size is assumed in advance. For example, if this is 9, a series of numbers from 1 to 9 is placed in table columns.
  • Below that, in a second row of tables, 9 randomly generated numbers for each interview are classified.
  • Before the interview begins, it is determined how many people in the respective household belong to the statistical population from which the sample is to be drawn (e.g. from the population of Germans entitled to vote in private households).
  • These people are listed according to their age.
  • If, for example, 6 possible respondents have now been identified in a household, first the number 6 and then the random number below it is searched for in the upper row of the table. Is this z. B. 3, then the third person on the list should be interviewed. However, if this person belonging to the household is not present or is otherwise unable to attend, later contact should be attempted for the interview with the third person.
Household size 1 2 3 4th 5 6th 7th 8th 9
Interviewee 1 1 2 4th 2 3 7th 4th 5

This example is a simple illustration of the principle of selection for people within households. In professional survey practice, further differentiations are used, for example, in that gender is also taken into account when creating household lists. There are therefore different implementation variants for the assignment of random numbers to possible respondents, but they are basically similar.

The household samples in ALLBUS 1980–1992, 1998 should be mentioned as a well-known, generally available reference survey in which the Swedish key method was used for the survey . In the case of a sample of persons (e.g. ALLBUS 1994, 1996, 2000–2008), the use of the Sweden key to select respondents in the households is not necessary or inappropriate, because there the respondents are subject to a fundamentally different selection process than those before are known for the first interview .

Other procedures

The birthday method works in a similar way, but is faster to perform, which is why it is mainly used for telephone interviews. However, it should not be seen as a variant of the Sweden key, but as an independent procedure. The person who was last born in the household is interviewed.

literature

  • Hans-Peter Kirschner: Sampling plan and weighting. In: Karl Ulrich Mayer , Peter Schmidt (Ed.): General population survey of the social sciences. Contributions to methodological problems of the ALLBUS 1980 (= monographs of social science methods. Vol. 5). Campus-Verlag, Frankfurt am Main et al. 1984, ISBN 3-593-33262-0 , pp. 114-182.
  • Leslie Kish: A Procedure for Objective Respondent Selection within the Household. In: Journal of the American Statistical Association. Vol. 44, No. 247 September 1949, ISSN  0162-1459 , pp. 380-387.