Traditional reliability analysis has been augmented with geometrical streng
th analysis using strain techniques, resulting in the conception of an exte
nsion to reliability theory called robustness analysis. To reflect contempo
rary statistical terminology, robustness is taken to mean insensitivity to
gross errors or blunders in the data. Robustness analysis is a natural, mer
ger of reliability and strain and is defined as the ability to resist defor
mations induced by the smallest detectable blunders as determined from inte
rnal reliability analysis. The geometrical strength analysis technique is u
sed in order to provide a more complete and detailed description of the pot
ential network deformation in terms of three independent measures represent
ing robustness in scale, orientation, and configuration. These measures are
also invariant with respect to "datum" shifts and orientation, and practic
ally invariant to changes in scale. Initial experiences with robustness ana
lysis have shown that it is a very powerful technique capable of providing
a detailed point-by-point assessment of the strength of a network.