An important feature of statistical matching is that the underlying joint distribution of the variables of interest is not identiﬁable. This produces a form on “uncertainty” on the statistical model. A measure to evaluate such an uncertainty is proposed. The effect of prior information in the form of constraints is exploited. Finally, the estimation of the proposed measure of uncertainty is studied.
Conti P., L., Marella, D. (2011). Measuring uncertainty in statistical matching. In CLADAG 2011- 8th Scientific Meeting of the Classification and Data Analysis Group of the italian statistical society.