REVIVAL OF A WELL-KNOWN, BUT IGNORED CORRECTION TERM FOR CHI(2) IN LINKAGE ANALYSIS WITH SMALL SAMPLES

The 'classical' chi(2) test is routinely applied in linkage analysis. The values customarily used, however, are only approximations since co ntinuous chi(2) terms are used instead of discrete multinomial terms. This error becomes increasingly important as the sample becomes small. In this note, a well-known correction term for improvement of this ap proximation is revived and numerically calculated for some interesting cases from linkage studies.