Weighting estimation under bipartite incidence graph sampling

Abstract Bipartite incidence graph sampling provides a unified representation of many situations for the purpose estimation, including existing unconventional methods, such as indirect, network or adaptive cluster sampling, which are not originally described problems. We develop large class design-based linear estimators, defined sample edges and subjected to general condition design unbiasedness. The contains special cases classic Horvitz-Thompson estimator, well other unbiased estimators in literature can be traced back Birnbaum et al. (1965). Our generalisation allows one devise future, thereby providing potential efficiency gains. Illustrations given line-intercept simulated graphs.