Ipsative data
In statistics , data is called ipsative (from the Latin ipse “self”) if the scales with which measurements are taken on different individuals differ from individual to individual. For example, a test subject is assessed using percentage and comparative estimates how strong a particular interest or motivation is compared to the other currently experienced interests; this comparison would be different for another test subject.
In particular, ipsative data is used when the different measured values obtained from an individual balance each other out in total. In statistics , ipsativization is also used to denote the centering (alignment) of several measured values recorded on the same individual on the individual overall mean .
Ipsative data are regularly generated through forced choice surveys in which the test subjects have to choose between mutually exclusive alternatives. An example of a forced choice question would be: “Which of the following two sentences best describes your interests: a.) I like arranging flowers. b.) I like to solve software problems. ”Both items are linked by the question, preference for flower arrangements correlates negatively with preference for software problems after the construction of the survey, although the preferences could correlate positively with another form of survey. This has consequences for the evaluation: Such individual-centered measurements can only be compared between people with great reservations. Ipsative data are mainly used when an intra-individual analysis of the properties is in the foreground. To be able to compare z. B. to be able to produce a norm group , you need normative data .
literature
- Theresa JB Kline: Psychological Testing: A Practical Approach to Design and Evaluation . Sage, 2005. ISBN 1412905443 , p. 70.
- Jürgen Bortz: Statistics for social scientists . Springer, Berlin 2005. ISBN 3-540-21271-X .
