A systematic quantification and classification of the vaguely known factors can be conducted by a sensitivity analysis. In what follows, a motivation for the so-called Distribution Based Global Sensitivity Analysis (DBGSA) is given. In particular, it is shown how the DBGSA can be calculated at low computational costs by combining Polynomial Chaos Expansion (PCE) and Point Estimate Method (PEM) principles. This seems to be a reasonable choice as PCE provides workable meta-models which can be efficiently parameterized by PEM-approximated multidimensional integrals.