[Skip to Navigation]
December 1997

What Is the 'Right' Statistical Measure of Twin Concordance (or Diagnostic Reliability and Validity)?

Arch Gen Psychiatry. 1997;54(12):1121-1124. doi:10.1001/archpsyc.1997.01830240081011

RECENT TWIN studies related to psychiatric disorders have used the tetrachoric correlation coefficient (TCC) as a statistical measure of twin concordance. While the TCC was introduced in 1901 by Karl Pearson,1-3 closely related in time and purpose with his development of the familiar product moment or Pearson correlation coefficient (PCC), over most of this century the TCC has been little known and little used in research applications. However, it has remained of continuing interest to statisticians and psychometricians.4-8 In great part its lack of use in applications may be due to its computational complexity, a problem now solved by ready access to compuers.9-11

See also page 1126  The TCC measures 2x2 correlation quite differently from more commonly used and familiar measures such as the ψ or K coefficients widely used in psychiatric research. Is the TCC the "right" measure to use to measure twin concordance? If not,