The Basic Four Measures and their Derivates in Dichotomous Diagnostic Tests

The paper focuses on four basic statistics of dichotomous diagnostic tests, i.e. sensitivity, specificity, positive and negative predictive value, and some of their derivates, like Youden index and predictive summary index, and on further derivates of these derivates, i.e. Matthews correlation coefficient (or Yule phi), chi squared test and Cramer's V coefficient. The paper contains also the necessary and sufficient conditions for a test to be invalid, to be uninformative and the necessary condition to be possibly valuable. The builder-concept of the paper is the determinant of 2 by 2 matrix.