Mostly Harmless Econometrics
由Joshua D. Angrist和Jörn-Steffen Pischke合作撰写的计量经济学经典著作Mostly Harmless Econometrics: An Empiricist's Companion详细介绍了应用实证研究中的核心计量工具，为社会科学研究者提供了一份精炼的操作指南。
然而，随机实验耗时长、成本高、可能遭遇学术伦理问题，因此对大多数学者来说未必具有现实可操作性。为此，本书作者以随机实验为基准（benchmark），把目光转向了自然实验（natural experiment）及准自然实验（quasi-experiment）。利用自然实验及准自然实验进行因果推断，需要充分利用本书所关注的核心计量工具：多元回归分析、工具变量方法（IV）和双重差分策略（DID）。本书在第三章主要讨论了多元回归分析方法。这一方法主要是指在控制了其他与残差项不相关的变量之后，用被解释变量对核心解释变量进行回归分析。该方法对提高估计准确性并揭示可能的因果关系大有裨益，而且也是接下来讨论的IV、DID等工具之基础。IV在本书第四章得到讨论。尽管工具变量不易寻找，但一旦找到合适的工具变量之后，使用两阶段普通最小二乘法便可较为精准地获得因果联系。当然，工具变量方法并非万无一失，局部有效性（LATE）等问题也受到了作者高度关注。DID在本书第五章得到呈现。作为处理遗漏变量问题、进行因果推论的有效方法，双重差分同样备受作者重视。与此相关，作者还在本章中就固定效应及面板数据处理进行了细致分析。以上便是本书的核心内容。接下来本书还进行了一些拓展讨论，主要涉及断点回归分析、分位数回归分析及回归分析中的标准差处理。
值得一提的是，不同于一般的计量经济学教科书，本书具有如下几方面有必要说明的特点：首先，本书并不对各种计量方法进行面面俱到的介绍，而是主要讨论在实证操作中处于核心地位的几类方法，对基本概念和技术问题的强调也穿插于核心方法的介绍及操作例证的讲解之中；其次，一般的计量经济学教科书非常关注经典假设及其违反的情况，本书则对此保持更为宽容的态度，并未在此花费太多篇幅；最后，在回归结果的统计性质中，本书更重视无偏性与一致性，对有效性的关注相对较弱。
评价：
“Finally – An econometrics book for practitioners! Not only for students, Mostly Harmless Econometricsis a fantastic resource for anyone who does empirical work.” — Sandra Black, UCLA
“This is a remarkable book–it does the profession a great service by taking knowledge that is usually acquired over many years and distilling it in such a succinct manner.” — Amitabh Chandra, Harvard Kennedy School of Government
“MHE is a fantastic book that should be read cover-to-cover by any young applied micro economist. The book provides an excellent mix of statistical detail, econometric intuition and practical instruction. The topic coverage includes the bulk of econometric tools used in the vast majority of applied microeconomics. I wish there was an econometric textbook this well done when I was in graduate school.” — Bill Evans, University of Notre Dame
Mostly Harmless Replication
一个大胆的尝试，用以下语言复制了《无害计量经济学》一书中的表格和数字: Stata R Python Julia
为什么要这么疯狂呢?我的主要动机是看看我是否可以在我的工作流中用R、Python或Julia替换Stata，所以我尝试用这些语言复制大部分无害的计量经济学。
目录如下：
Questions about Questions
The Experimental Ideal
Making Regression Make Sense
Instrumental Variables inAction
Parallel Worlds
Getting a Little Jumpy
Quantile Regression
Nonstandard Standard Error Issues
查看Wiki中的入门指南，了解使用每种语言设置机器的技巧。
代码链接 https://gitee.com/econometric/mostly-harmless-replication
精通计量Mastering 'Metrics一书数据及代码
本页面链接了许多Mastering 'Metrics实证案例背后的研究和数据集。
文中链接失效，详情可以进入原文查看http://www.masteringmetrics.com/resources/
Chapter 1: Table 1.1
This table compares people with and without health insurance in the 2009 National Health Interview Survey (NHIS).
Studies
The 2009 NHIS
Data
ReadMe
Data
.do files
These tables report our analysis of data from the RAND Health Insurance Experiment (HIE).
Studies
The RAND HIE
Brook (1983)
Aron-Dine, Einav, and Finkelstein (2013)
Data
ReadMe
Data
.do files
These tables report results from the Oregon Health Plan (OHP) studies.
Studies
Baicker et al. (NEJM 2013)
Finkelstein et al. (QJE 2012)
Taubman et al. (Science 2014)
Data
ReadMe
Data
This table presents the distribution of assigned and delivered treatments from the Minneapolis Domestic Violence Experiment (MDVE).
Studies
Sherman and Berk (1984)
Angrist (2006)
Data
ReadMe
Data
.do file
These contain our RD analysis of the minimum legal drinking age (MLDA).
Studies
Carpenter and Dobkin (2009)
Carpenter and Dobkin (2011)
Data
ReadMe
Data
.do files
These figures show a DD analysis of the effect of monetary policy on bank failures in Mississippi in 1930.
Studies
Richardson and Troost (2009)
Data
ReadMe
Data
.do file
These tables report our DD analysis of the MLDA.
Studies
Du Mouchel, Williams, and Zador (1987)
Norberg, Bierut and Grucza (2009)
Data
ReadMe
Data
.do files
This table reports estimates of the returns to schooling for Twinsburg twins.
Studies
Ashenfelter and Krueger (1994)
Ashenfelter and Rouse (1998)
Data
ReadMe
Data
.do file
This table reports 2SLS estimates of the returns to schooling using child labor laws as instruments for years of schooling.
Studies
Acemoglu and Angrist (2000)
Data
ReadMe
Data
.do file
These tables and figures present an IV analysis of the returns to schooling using quarters of birth (QOB) as instruments for years of schooling.
Studies
Angrist and Krueger (1991)
Data
ReadMe
Data
.do file
These figures analyze the “sheepskin effects” of a high school diploma.
Studies
Clark and Martorell (2014)
Data
ReadMe
Data
.do file返回搜狐，查看更多
责任编辑：