National Medical Expenditure Survey, 1987: Survey of American Indians and Alaska Natives, Preliminary Health Status Questionnaires and Access to Care Supplement [Public Use Tape 21P]

The sample is a stratified area probability design with three stages of sample selection: (1) selection of primary sampling units (PSUs), which are counties or groups of contiguous counties (in Alaska, the county equivalents developed by the state of Alaska and the Census Bureau for statistical purposes were employed), (2) selection of segments within PSUs, (3) selection and screening of dwelling units within segments. Based on the results of the screening, all dwelling units including persons eligible for IHS services were selected in the SAIAN. The sample frame included all counties in the United States including or adjacent to reservations of federally-recognized tribes or Alaska Native villages. For cost considerations, the sample frame was truncated to remove counties with the lowest concentrations of eligible persons, and disproportionate sampling was applied to remaining counties with low concentrations of American Indians.本样本采用分层区域概率抽样设计，共设置三阶段抽样流程：（1）抽取初级抽样单元（primary sampling units, PSUs），即单个县或连片县群（阿拉斯加州采用该州与美国人口普查局为统计用途划定的县级等效区域）；（2）在初级抽样单元内抽取街区段；（3）在街区段内选取并筛查住宅单元。基于筛查结果，在SAIAN项目中，所有包含符合IHS服务（Indian Health Service）资格人员的住宅单元均被纳入样本。本次抽样框涵盖美国境内所有毗邻或坐落于联邦认可部落保留地、阿拉斯加原住民村落范围内的县。出于成本控制考量，抽样框进行了截断处理，剔除了合格人员集中度最低的县，并针对剩余印第安人集中度较低的县采用了不等比例抽样策略。