A Successive Differences Method for Growth Curves with Missing Data and Random Observation Times
提出逐次差分法的推广形式，通过划分时间子区间处理观测时间不固定的情况，并给出平行性假设检验，应用于血清胆固醇重复测量数据。
Abstract Incomplete growth curve data can be analyzed by the successive differences (SD) method, which uses the difference in consecutive pairs of observations for all subjects having two or more repeated measures. We have generalized it to handle varying observation times as well by partitioning the time interval spanned by all repeated measures into subintervals. The model is developed, including test statistics for the hypotheses of parallelism and of no change in response level over time, assuming parallelism. This generalized SD method is then applied to repeated serum cholesterol measurements on a subsample of 1,072 men from a prospective cohort study of 8,006 Hawaiian Japanese men living on Oahu.