REGRESSION TO THE MEAN

被引：19

作者：

NEWELL, D ^{[1
]}

SIMPSON, J ^{[1
]}

机构：

[1] UNIV SYDNEY,DEPT PUBL HLTH,SYDNEY,NSW 2006,AUSTRALIA

来源：

MEDICAL JOURNAL OF AUSTRALIA | 1990年 / 153卷 / 03期

关键词：

D O I：

10.5694/j.1326-5377.1990.tb136839.x

中图分类号：

R5 [内科学];

学科分类号：

1002 ; 100201 ;

摘要：

An 'obvious' clinical improvement in patients with an extreme initial reading on any biological measurement may be partially if not totally explained by an old-established but not always recognized statistical phenomenon. This same phenomenon of regression to the mean can also lead to a mistaken belief that response to a treatment is best in patients with the worst initial readings. It is best to plan studies which avoid the effects of regression to the mean. If this is not done statistical methods should be used to adjust for it.

引用

页码：166 / 168

页数：3

共 14 条

[1] RELATION BETWEEN CHANGE AND INITIAL VALUE [J].