HISPANIC COMMUNITY HEALTH STUDY - STUDY OF LATINOS COORDINATING CENTER DIVERSITY SUPPLEMENT FOR JOSE LOPEZ

磨损经常发生在纵向研究中，例如，参与者退出研究，导致数据缺失。有三种常见的数据丢失机制：完全随机丢失(MCAR)、随机丢失(MAR)和不随机丢失(MNAR)。众所周知，如果返回后续访问的参与者与没有进行后续访问的参与者在系统上不同，那么只对那些返回的参与者使用标准方法可能会提供有偏见的结果。正确计算丢失的数据对于获得有效的结果至关重要。调整单元格权重(ACW)是在缺失是随机的情况下调整基线抽样权重(BSW)的一种方法。这一多样性补充项目建议在使用3级或更多级别的MLMS分析复杂的纵向调查数据时实施ACW。

Attrition often occurs in longitudinal studies, e.g., participants dropped out of the study, and results in missing data. There are three general categories of missing data mechanisms: missing completely at random (MCAR), missing at random (MAR), and not missing at random (MNAR). It is well known that if participants who returned to follow-up visits are systematically different from those who did not, using standard methods on only those who returned could provide biased results. Properly accounting for the missing data is critical to obtain valid results. Adjustment Cell Weighting (ACW) is one approach to adjust Baseline Sampling Weights (BSW) when missing is at random. This diversity supplement project proposes to implement ACW when analyzing complex longitudinal survey data using MLMs with 3 or more levels.