A unified framework is proposed to provide a comprehensive understanding of the problem of data association in multitarget tracking. Under this framework, a systematic scheme is developed for generating the data association hypotheses in the target-oriented, measurement-oriented, and track-oriented approaches. Since there are many data association hypotheses with identical likelihoods in the measurement-oriented and track-oriented approaches, two specialized algorithms are developed to efficiently generate the data association hypotheses in the two approaches by adapting the depth-first search (DFS) algorithm developed earlier in the target-oriented case.