Nearly random designs with greatly improved balance

Nearly random designs with greatly improved balance
复制标题

近乎随机的设计,平衡性大大提高

DOI:
10.1093/biomet/asz026
复制
发表时间:
2016
期刊:
影响因子:
2.7
通讯作者:
A. Kapelner
A. Kapelner
中科院分区:
数学2区
文献类型:
--
作者:
A. Krieger;David Azriel;A. Kapelner

文献摘要

参考文献

被引文献

相似文献

我们提出了一个程序,将一组实验单位分成两组,这两组在预先指定的协变量集上是相似的,并且几乎与完全随机化一样随机。在完全随机化的情况下,治疗组和对照组之间的标准化平均值的差异为O_{\rmp}(n^{-1/2}),这在小样本中可能是重要的。我们提出了一个算法,减少不平衡为O_{\rmp}(n^{-3})$为一个协变量和O_{\rmp}\{n^{-(1 + 2/p)}\}$为$p$协变量,但其分配,严格地说,非随机的。除了度量的最大特征值的分配方差,我们引入了两个度量,捕获偏离随机化,并表明我们的分配几乎是随机的完全随机化的所有措施。模拟说明的结果,并推断进行了讨论。一个R软件包可以根据我们的算法和其他流行的设计生成设计。
We present a procedure that divides a set of experimental units into two groups that are similar on a prespecified set of covariates and are almost as random as with a complete randomization. Under complete randomization, the difference in the standardized average between treatment and control is $O_{\rm p}(n^{-1/2})$, which may be material in small samples. We present an algorithm that reduces imbalance to $O_{\rm p}(n^{-3})$ for one covariate and $O_{\rm p}\{n^{-(1 + 2/p)}\}$ for $p$ covariates, but whose assignments are, strictly speaking, nonrandom. In addition to the metric of maximum eigenvalue of allocation variance, we introduce two metrics that capture departures from randomization and show that our assignments are nearly as random as complete randomization in terms of all measures. Simulations illustrate the results, and inference is discussed. An R package to generate designs according to our algorithm and other popular designs is available.
DOI: 10.1016/0197-2456(88)90046-3
发表时间: 1988-12-01
期刊: CONTROLLED CLINICAL TRIALS
影响因子: --
作者:
LACHIN, JM
通讯作者: LACHIN, JM
DOI: 10.1093/biostatistics/5.2.263
发表时间: 2004-04-01
期刊: BIOSTATISTICS
影响因子: 2.1
作者:
Greevy, R;Lu, B;Rosenbaum, P
通讯作者: Rosenbaum, P
DOI: 10.1073/pnas.1808191115
发表时间: 2018-09-11
影响因子: 11.1
作者:
Li, Xinran;Ding, Peng;Rubin, Donald B.
通讯作者: Rubin, Donald B.