Projection (statistics)
This item has been on the quality assurance side of the portal sociology entered. This is done in order to bring the quality of the articles on the subject of sociology to an acceptable level. Help eliminate the shortcomings in this article and participate in the discussion . ( Enter article )
Reason: Sources are unsuitable. - Zulu55 ( discussion ) 11:48, Aug 27, 2012 (CEST)
In statistical extrapolation, a projection is a method of combining different statistically determined scenarios (at least two) with one another and thus either depicting reality more precisely than the determination of individual studies could provide, or gaining additional knowledge that can be derived from the analysis individual data would not result. The theoretical basis is formed by the theories that every statistically determined extrapolation contains a margin of error that can be verified by combining it with other statistical values and thus reduced, and that related scenario values can be constructed from the combination of several statistical values.
Example election research
The classic scenarios are an election survey in the past and the actual voting behavior in temporal proximity to one another. Scenario 1 : the actual voting behavior can be predicted with a few thousand voters. Scenario 2 : Some voters from a smaller party vote for a large party in the actual election . They prefer the government of a large party, but actually elect a small party that is supposed to form a coalition with the large party ( error rate : tactical voting behavior) Projection : The error rate determined can now be used in future surveys, which are comparable in type and scope with scenario 1 , project, d. H. the statistically determined values of a current survey are corrected by the values of the error rate already determined in the assumption that the presumed reality can thereby be mapped even more precisely. Another scenario can arise when voters of extreme parties do not admit this in a poll and tick another party. Other factors can be included in the projection, e.g. B. the change in the spectrum of parties, the influence of the currently discussed political issues on party competence, etc.
Example market research
In a market analysis , values can be determined for which the created database is incomplete and has to be expanded to include scenarios of other data. Example OpenOffice.org : A market study was carried out among customers who use the OpenOffice.org software in their company ( scenario 1 , own market research ). In this way, general statistical values about migration and operating time can be determined in relation to the company or workplace. If one now wants to determine the share of this software in the entire market of office programs in the professional environment, one must also determine the relationship between privately used and officer-relevant company PCs ( scenario 2 , third-party, non-program-specific data) and the overall distribution of the program ( scenario 3 , Numbers above the program paragraph). The projection now projects the distribution (scenario 3) onto the share of company computers (scenario 2) and can use this to determine the projected market share of the software in the professional area and thus supplement the program-related market study with the market share of the program.
swell
- ↑ Projection at the research group Wahlen
- ↑ Study ( Memento of the original from December 5, 2008 in the Internet Archive ) Info: The archive link was inserted automatically and has not yet been checked. Please check the original and archive link according to the instructions and then remove this notice. (PDF; 95 kB) from MIC Consulting, Wiesbaden (Thomas Krumbein)
