The Census Bureau uses a 'capture-recapture' or dual system estimation (DSE) methodology to es­timate total population including those missed by the census. The two 'systems' are the census and a Post Enumeration Survey (PES). One of the assumptions underlying use of the DSE to estimate population size is that within each poststratum (defined by some set of geographic and demographic variables), being in the census is inde­pendent of being in the PES. When these events are not independent, there is a 'correlation bias' which typically leads to underestimation of the number of people who are in neither the census nor the PES. Reasons for the possible failure of this assumption of independence have been discussed (Hogan and Wolter 1988). One method of checking this assump­tion, or indeed of measuring the correlation bias, makes use of a third source of names and addresses - an alternative list. In this paper, we discuss methods of estimat­ing the number of movers and non-movers, cross­classified by inclusion in census, PES, and alterna­tive list (administrative or ethnographer's lists). We also discuss how these estimates can be used to give total population estimates, and the relative merits of each estimate.