摘要
This paper introduces a measure which obviates the need to consult any tables.Average value x’ and standard deviation s’in this measure arc computed without ineluding suspend value xm.If│xm一x’│/s’>2.58.then xm, is an outlying obscrvation;when n≥9,fiducial probability p>95一99%.As the standard deviation of population σ is given,if│xm一x’│/σ>2.58.tht xm is outlying;when n≥4,p>97一99%.This article also demonstrates that comparing with measures of Grubbs,Nair and Chauvenet.ours is more efficient in judgement under the same or large p.
This paper introduces a measure which obviates the need to consult any tables.Average value x' and standard deviation s’in this measure arc computed without ineluding suspend value xm.If│xm一x'│/s’>2.58.then xm, is an outlying obscrvation;when n≥9,fiducial probability p>95一99%.As the standard deviation of population σ is given,if│xm一x’│/σ>2.58.tht xm is outlying;when n≥4,p>97一99%.This article also demonstrates that comparing with measures of Grubbs,Nair and Chauvenet.ours is more efficient in judgement under the same or large p.
出处
《统计研究》
CSSCI
北大核心
1996年第4期65-68,共4页
Statistical Research
