What does validity mean? Definition and example

Every empirical research should ideally measure exactly what is captured in the research question. If this is not the case, then the results of the investigation are also not meaningful. With the quality criterion of validity, science therefore tests the substantive validity of the measurement.

General definition of validity

Validity (= correctness or accuracy of the measurement) reflects the extent to which a research work in its entirety actually achieves the results that correspond to the stated research objective.

Validity as an essential evaluation criterion of every scientific work is considered fulfilled when the research question is adequately addressed and answered with respect to the chosen topic. In other words: if what is measured in the investigation is actually what should be measured.

Example of validity:

If in a statistics exam only one-third of the questions relate to the course material covered by students over the semester, because, for example, an old exam from the previous course was used, then the exam is not actually valid in terms of content. The test instrument (= the exam with outdated questions) does not capture what it should capture in this case (= the covered material from the current lecture) and validity would not be given in this case.

 

 

Scientific testing of validity

In scientific practice, the investigation is considered valid if its components – such as the concept or operationalization – are substantively valid. The following options for testing validity are common:

1. Content validity
2. Criterion validity
3. Construct validity
4. Expert validity
5. Known-Group Validity

 

Content validity

(= Substantive completeness of the investigation): Content validity examines whether all units of the study are substantively well-thought-out, e.g. whether a questionnaire within a quantitative survey is meaningfully and logically structured. It can be justified relatively easily through preliminary considerations about the structure of the research.

 

Criterion validity

(= Comparison of collected data to another measurement criterion): Criterion validity examines whether the achieved results could also be obtained with a different measurement instrument. In a quantitative survey on the social media behavior of adolescents, higher criterion validity could be achieved through an additional content analysis of technical articles on this topic.

 

Construct validity

(= Fundamental appropriateness of the entire research design): Construct validity exists when the measurement instrument used, e.g. a survey, can not only be successfully applied in practice, but also establishes a theoretical connection to other important research in the same field.

 

Expert validity

(= Comparison of collected data with the opinion of an expert): The expert validity of a study on news consumption could be tested or confirmed by a renowned scientist in the field of television research.

 

Known-Group Validity

(= Comparison of collected data with a selected group of test persons): Here, people with an extreme opinion on the topic being investigated are surveyed, e.g. pro-gamers on the dangers of first-person shooter games. If the mean of the known group differs significantly from that of the group actually being investigated (= e.g. the population average), the known-group validity of the study would be considered confirmed

 

When assessing validity, it is simply impossible to meet all the criteria presented here. For student work in particular, the limitation to content and construct validity is usually completely sufficient.