在现场流行病学调查中，大多采用分阶段随机抽样的方法获取数据，如先随机抽取若干个省，再抽取省下面的县，县下面的乡，乡下面的村，直到家庭或个体等。显然这样的数据具有层次结构性，而同一层次下的个体具有层内相似性。以前处理这样的数据通常忽略层次结构或者分层（如按 In field epidemiological surveys, most of the data were collected using a phased random sampling method. For example, a number of provinces were selected at random, and the provinces under the counties and counties below the county, and villages below the township were sampled until families or individuals. Obviously, such data has a hierarchical structure, and individuals at the same level have intra-layer similarities. Previously dealing with such data usually ignores hierarchies or layers (eg