Consider three different but related problems with auxiliary information: infinite population sampling or Monte Carlo with control variates, missing response with explanatory variables, and Poisson and rejective sampling with auxiliary variables. We demonstrate unified regression and likelihood...