This talk describes methods for masking public-use microdata that are intended to preserve analytic validity and assure that individual microdata cannot be re-identified. The additive noise methods generalize work of Kim (1986), Fuller (1993), and Tendick and Matloff (1994) and appeared in the 2002 Springer LNCS Monograph "Inference Control in Statistical Databases" (J. Domingo, Ed.). We also give some relationships to other recent work (IEEE TIKE: Samarati 2001, Domingo-Ferrer and Mateo-Sanz 2002; SIGKDD: Iyengar 2002).