Differential item functioning
Differential item functioning (DIF) or group-dependent indicator is a term from psychological diagnostics that arose from the use of psychological instruments in different cultures or subcultures and is intended to describe different response behavior in groups to the same questions. Group dependency arises when people from different groups (mostly gender or ethnicity ) with the same personality traits or with the same ability have a different probability of giving a certain answer to a questionnaire or a test.
The analysis of the group dependency enables an indication of unexpected behavior in relation to an indicator ( item ). An indicator does not yet show any group dependency if people from different groups have different response probabilities, on the other hand it shows group dependency when people from different groups with the same underlying characteristics show a different probability of giving a certain answer.
The analysis and correction of group dependency plays a particularly important role in so-called high-stake tests , in which not only a property is measured on the basis of interest, but a selection decision is made across different linguistic areas . B. for study place management in multilingual countries such as Switzerland , Canada , Italy etc.
example
In a nationwide knowledge test among schoolchildren, the following question is used to ask about their geographical knowledge: “What is the capital of Hesse?” Hessian schoolchildren have an advantage because they are much more likely to know the capital of their own federal state. The indicator is therefore group-dependent and the test result would falsely give the impression that Hessian students have better geographical knowledge than students from other federal states.
Individual evidence
- ↑ Embretson, SE, travel, SP (2000). Item Response Theory for Psychologists.