Loading Tree…
Different data values appear in seemingly erroneous combinations.
The indicator uncertain contradictions is used to identify combinations of data values that may not be entirely impossible but unlikely. For example gender is most often a stable characteristic of a person. However, changes in gender identity may occur.
Examples of uncertain contradictions are:
a non-smoker will usually not buy tobacco products
if eating preference is vegetarian or vegan, weekly meat consumption should be zero.
Any empirical contradictions implies an elevated probability of some data quality issue which requires further investigation. Should no such check be possible an elevated count of uncertain contradictions can be interpreted as an indication of a lower data quality.
The higher the number or percentage of uncertain contradictions the potentially lower the data quality.
Nonnemacher M, Nasseh D, Stausberg J. Datenqualität in der medizinischen Forschung: Leitlinie zum Adaptiven Datenmanagement in Kohortenstudien und Registern. Berlin: TMF e.V..; 2014.
Stausberg J, Bauer U, Nasseh D, et al. Indicators of data quality: review and requirements from the perspective of networked medical research MIBE 2019;15(1):1-8.