可忽略性

此词条暂由彩云小译翻译，翻译字数共687，未经人工整理和审校，带来阅读不便，请见谅。

In statistics, ignorability is a feature of an experiment design whereby the method of data collection (and the nature of missing data) do not depend on the missing data. A missing data mechanism such as a treatment assignment or survey sampling strategy is "ignorable" if the missing data matrix, which indicates which variables are observed or missing, is independent of the missing data conditional on the observed data.

在统计学中，可忽略性是一种实验设计特征，即数据收集方法(以及缺失数据的性质)不依赖于缺失数据。若显示哪些变量已观测或缺失的缺失数据矩阵与已观测数据为条件的缺失数据相互独立，则称该数据缺失机制（例如处理分配或抽样调查策略）是“可忽略的”。

This idea is part of the Rubin Causal Inference Model, developed by Donald Rubin in collaboration with Paul Rosenbaum in the early 1970s. The exact definition differs between their articles in that period. In one of Rubins articles from 1978 Rubin discuss ignorable assignment mechanisms, which can be understood as the way individuals are assigned to treatment groups is irrelevant for the data analysis, given everything that is recorded about that individual. Later, in 1983 Rubin and Rosenbaum rather define strongly ignorable treatment assignment which is a stronger condition, mathematically formulated as [math]\displaystyle{ (r_1,r_0) \perp \!\!\!\perp z \mid v ,\quad 0\lt \operatorname{pr}(z=1)\lt 1 \quad \forall v }[/math], where [math]\displaystyle{ r_t }[/math] is a potential outcome given treatment [math]\displaystyle{ t }[/math], [math]\displaystyle{ v }[/math] is some covariates and [math]\displaystyle{ z }[/math] is the actual treatment.

这个想法是20世纪70年代早期Donald Rubin和Paul Rosenbaum 合作提出的鲁宾因果推理模型的一部分。但那时，他们文章中可忽略性的确切定义不同。1978年鲁宾在一篇文章中讨论了可忽略的分配机制, 其可理解为将个体分配到处理组的方式与数据分析无关，因为已经记录了有关该个体的所有信息。后来，在 1983 年，Rubin 和 Rosenbaum 更确切地定义了“处理分配的强可忽略性”， ，这是一个更强的假设条件，数学公式为[math]\displaystyle{ (r_1,r_0) \perp \!\!\!\perp z \mid v ,\quad 0\lt \operatorname{pr}(z=1)\lt 1 \quad \forall v }[/math]，其中[math]\displaystyle{ r_t }[/math]是给定处理状态 [math]\displaystyle{ t }[/math]下的潜在结果，[math]\displaystyle{ v }[/math] 是协变量，[math]\displaystyle{ z }[/math] 是实际的处理状态。

Pearl [2000] devised a simple graphical criterion, called back-door, that entails ignorability and identifies sets of covariates that achieve this condition.

Pearl [2000]设计了一个简单的图形标准，称为“后门”(back-door) ，它需要可忽略性并确定达到这种条件的协变量集。

Ignorability (better called exogeneity) simply means we can ignore how one ended up in one vs. the other group (‘treated’ Tx = 1, or ‘control’ Tx = 0) when it comes to the potential outcome (say Y). It was also called unconfoundedness, selection on the observables, or no omitted variable bias.

可忽略性(称为外生性更好)其简明含义是，当涉及到潜在结果（Y）时，一个人是怎样最终处于一个群体中而非另一个群体中(“处理组”Tx = 1，或“控制组”Tx = 0)我们是可忽略的。它也被称为非混淆性，基于可观测变量的选择选择的可观察的，或无遗漏变量偏差。

Formally it has been written as [Y_i1, Y_i0] ⊥ Tx_i, or in words the potential Y outcome of person i had they been treated or not does not depend on whether they have really been (observable) treated or not. We can ignore in other words how people ended up in one vs. the other condition, and treat their potential outcomes as exchangeable. While this seems thick, it becomes clear if we add subscripts for the ‘realized’ and superscripts for the ‘ideal’ (potential) worlds (notation suggested by David Freedman; a visual can help here: potential outcomes simplified).

数学形式上，它被写成[math]\displaystyle{ [Y\lt sub\gt i\lt /sub\gt 1, Y\lt sub\gt i\lt /sub\gt 0] ⊥ Tx\lt sub\gt i\lt /sub\gt }[/math] ，或者用文字来说，人们的潜在结果Y我已经治疗或不治疗不取决于他们是否真的被(可观察的)治疗。换句话说，我们可以忽略人们是如何在一种情况下和另一种情况下结束生命的，而把他们的潜在结果看作是可以交换的。虽然这看起来很厚，但是如果我们为“理想”(潜在)世界添加“已实现”的下标和上标就变得很清楚了(由 David Freedman提出的符号; 一个视觉可以在这里帮助:potential outcomes simplified).

So: Y₁¹/*Y₀¹ are potential Y outcomes had the person been treated (superscript ¹), when in reality they have actually been (Y₁¹, subscript ₁), or not (*Y₀¹: the * signals this quantity can never be realized or observed, or is fully contrary-to-fact or counterfactual, CF).

所以:Y₁¹/*Y₀¹是潜在结果Y，如果个体被处理(superscript ¹) ，那么实际上它们是(Y₁¹, subscript ₁) ，而不是(*Y₀¹:: * 表示这个值是无法实现或不可观测的，或完全与事实或事实相反，CF)。

Similarly, *Y₁⁰/Y₀⁰ are potential Y outcomes had the person not been treated (superscript ⁰), when in reality they have been (*Y₁⁰, subscript ₁), or not actually (Y₀⁰).

同样，*Y₁⁰/Y₀⁰是个体未被处理 (上角标0)的潜在结果Y，当现实中它们是(*Y₁⁰, subscript ₁),或实际上不是 (Y₀⁰).

Only one of each potential outcome (PO) can be realized, the other cannot, for the same assignment to condition, so when we try to estimate treatment effects, we need something to replace the fully contrary-to-fact ones with observables (or estimate them). When ignorability/exogeneity holds, like when people are randomized to be treated or not, we can ‘replace’ *Y₀¹ with its observable counterpart Y₁¹, and *Y₁⁰ with its observable counterpart Y₀⁰, not at the individual level Y_i’s, but when it comes to averages like E[Y_i¹ – Y_i⁰], which is exactly the causal treatment effect (TE) one tries to recover.

对于相同的条件分配，每个潜在结果(PO)中只有一个是实际发生的，而另一个不会发生，因此当我们试图估计治疗效果时，我们需要用可观测值(或估计值)来代替完全相反的结果。当可忽略性/外生性成立时，如人们被随机分配治疗与否，我们可以用可观察到的*Y₀¹’替换‘Y₁¹,而 *Y₁⁰ 与其对应的Y₀⁰，不在 Y_i的个体水平，而是 E[Y_i¹ – Y_i⁰]平均层面而言在 ，这样的平均值时，正是因果治疗效应(TE)试图恢复的结果。

Because of the ‘consistency rule’, the potential outcomes are the values actually realized, so we can write Y_i⁰ = Y_i0⁰ and Y_i¹ = Y_i1¹ (“the consistency rule states that an individual’s potential outcome under a hypothetical condition that happened to materialize is precisely the outcome experienced by that individual”, p. 872). Hence TE = E[Y_i¹ – Y_i⁰] = E[Y_i1¹ – Y_i0⁰].

由于“一致性规则”，潜在的结果是实际实现的价值，因此我们可以写 Y_i⁰ = Y_i0⁰ and Y_i¹ = Y_i1¹("一致性规则指出，假设个体在某种条件下实现的潜在结果恰恰是该个体所经历的结果", p. 872).因此，TE = E[Y_i¹ – Y_i⁰] = E[Y_i1¹ – Y_i0⁰].

Now, by simply adding and subtracting the same fully counterfactual quantity *Y₁⁰ we get:

现在，通过简单的加减相同的完全反事实量 *Y₁⁰ 我们得到:

E[Y_i1¹ – Y_i0⁰] = E[Y_i1¹ –*Y₁⁰ +*Y₁⁰ - Y_i0⁰] = E[Y_i1¹ –*Y₁⁰] + E[*Y₁⁰ - Y_i0⁰] = ATT + {Selection Bias},

E[Y_i1¹ – Y_i0⁰] = E[Y_i1¹ –*Y₁⁰ +*Y₁⁰ - Y_i0⁰] = E[Y_i1¹ –*Y₁⁰] + E[*Y₁⁰ - Y_i0⁰] = ATT + {Selection Bias},

where ATT = average treatment effect on the treated and the second term is the bias introduced when people have the choice to belong to either the ‘treated’ or the ‘control’ group.

其中 ATT = 处理组的平均处理效应，第二项是个体选择属于处理组或对照组时引入的偏差。

Ignorability, either plain or conditional on some other variables, implies that such selection bias can be ignored, so one can recover (or estimate) the causal effect.

可忽略性，无论是普通的还是条件性的，都意味着这种选择偏差可以被忽略，因此人们可以恢复(或估计)因果效应。

See also

• 随机缺失

推荐阅读

• Gelman, Andrew; Carlin, John B.; Stern, Hal S.; Rubin, Donald B. (2004). Bayesian Data Analysis. New York: Chapman & Hall/CRC. 

• Pearl, Judea (2000). Causality: Models, Reasoning, and Inference. Cambridge University Press. https://archive.org/details/causalitymodelsr0000pear. 

• Jaeger, Manfred (2011). "Ignorability in Statistical and Probabilistic Inference". Journal of Artificial Intelligence Research. 24: 889–917. arXiv:1109.2143. Bibcode:2011arXiv1109.2143J. doi:10.1613/jair.1657.

Category:Design of experiments

类别: 实验设计

Category:Causal inference

类别: 因果推理

This page was moved from wikipedia:en:Ignorability. Its edit history can be viewed at 可忽略性/edithistory