作者: Robert. W. Mee
DOI: 10.1080/03610928808829692
关键词:
摘要: Point and interval estimation for the percentage of a normal population lying outside specified Interval Is investigated. Resnikoff's (1955) confidence procedure is shown to perform well when small, be conservative otherwise.When estimated large, an alternative recommended.Monte Carlo studies show sensitivity these procedures violations normality assumption.Attempts transform sample approximate prior were moderately successful in achieving robust procedures.