Confidence limits for the abscissa of intersection of two linear regressions

Abstract

A method to determine the confidence limits for the abscissa of the intersection of two linear regressions has been developed. This method does not require the assumption of equal variance for the two regressions, as was necessary with previous methods. A numerical example is included on thermodynamic, glass transition data for which this method is applicable. Comparisons are made between the results using equal and unequal variance assumptions. A FORTRAN subroutine is included for computations using both assumptions.