Screening procedures aim at identifying persons who have a certain (binary) characteristic. They are often used to identify people at risk who later develop mental disorders, diseases or school problems. For the evaluation of the screenings there are a number of characteristic values, of which sensitivity and specificity are the best known. They show how many people were correctly identified by the test as risk and how many as non-risk persons.
Besides these basic data, however, there are a number of other important data, such as the positive and negative predictive values or the Matthews Correlation Coefficient (MCC, relative increase in the hit rate compared to the random hit rate; further information and a detailed description of the various parameters can be found in Marx & Lenhard, 2010).
You can use the table to determine various characteristic values for screening procedures. Please enter the absolute case numbers in the table cells.
Area proportions see graphic please fill in absolute numbers |
Predictor (e. g. from a screening) |
||
positive result | negative result | ||
Criteria | not affected | b (false positive) |
d (correct negative) |
affected | a (correct positive) |
c (false negative) |
Quality indicators of the screening: | |
Sensitivity | |
Specificity | |
positive predictive value | |
negative predictive value | |
Hit ratio | |
Random hit ratio | |
Matthews correlation coefficient (MCC) |
Explanation:
The online tool is available as an Excel sheet and as Source-Code in Java under the General Public License (GPL):
Citeable: Lenhard, W. & Lenhard, A. (2014). Calculation of test quality criteria for screenings. available: http://www.psychometrica.de/testkennwerte.html. Dettelbach: Psychometrica.