Matching as Nonparametric Preprocessing for Reducing Model Dependence in Parametric Causal Inference

By Daniel E. Ho, Kosuke Imai, Gary King, Elizabeth A. Stuart. Political Analysis, 2007

4,316 citations recorded by the source corpus.

Although published works rarely include causal estimates from more than a few model specifications, authors usually choose the presented estimates from numerous trial runs readers never see. Given the often large variation in estimates across choices of control variables, functional forms, and other modeling assumptions, how can researchers ensure that the few estimates presented are accurate or representative? How do readers know that publications are not merely demonstrations that it is possible to find a specification that fits the author's favorite hypothesis? And how do we evaluate or even define statistical properties like unbiasedness or mean squared error when no unique model or estimator even exists? Matching methods, which offer the promise of causal inference with fewer assumptions, constitute one possible way forward, but crucial results in this fast-growing methodological literature are often grossly misinterpreted. We explain how to avoid these misinterpretations and propose a unified approach that makes it possible for researchers to preprocess data with matching (such as with the easy-to-use software we offer) and then to apply the best parametric techniques they would have used anyway. This procedure makes parametric models produce more accurate and considerably less model-dependent causal inferences.

Subfields: Statistics and Probability

View Paper: https://www.semanticscholar.org/paper/d1150b048aa90d948aa0f0b56da661ca3c99737f

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Statistics and Probability

Citations in the corpus

Chief Executive Officer Equity Incentives and Accounting Irregularities (2010, positive): “…achieve adequate covariate balance, an alternative approach is to form matched pairs with the propensity scores and then estimate a (conditional logistic) regression of the outcome as a function of treatment and the vector of control variables used in the propensity-score model (Ho et al., 2007).”
Co-Curricular Connections: The Role of Undergraduate Research Experiences in Promoting Engineering Students’ Communication, Teamwork, and Leadership Skills (2016, methodology): “5 % is seen as sufficient (Ho et al., 2007).”
Bicycle Infrastructure and Traffic Congestion: Evidence from DC's Capital Bikeshare (2018, positive): “…scores, P (X), which denotes the predicted probability that a particular block group will have a bikeshare dock conditional on X. Note that we combine the identifying power of the FEPD estimator with that of matching, following the approach of Ho et al. (2007) and Ferraro and Miranda (2014).”
Impact of Increased Academic Intensity on Transfer Rates: An Application of Matching Estimators to Student-Unit Record Data (2009, methodology): “Estimation was completed using the Match It library for the R statistical computing environment, as described in Ho et al. (2005). As would be expected even given the small number of variables available, the number of subclasses produced by such a procedure is relatively large.”
Putting PjBL to the Test: The Impact of Project-Based Learning on Second Graders’ Social Studies and Literacy Learning and Motivation in Low-SES School Settings (2020, neutral): “In line with relevant methodological research (e.g., Ho et al., 2007), the Institute of Education Sciences What Works Clearinghouse (2014) considers an ES of 0.25 or higher to be ‘‘substantively important’’ (p. 23).”
Analyzing Individual and Average Causal Effects via Structural Equation Models (2005, background): “In this paper it has been argued that Holland’s (1986) fundamental problem of causal inference is not ubiquitous, and that there are phenomena in Psychology and related disciplines for which it is possible to use subjects as their own control in pretests.”
Protecting Ecosystems and Alleviating Poverty with Parks and Reserves: ‘Win-Win’ or Tradeoffs? (2011, result): “[14]”
Risk in Islamic banking and corporate governance (2018, background): “…sample in a comparative research setting may cause sample selection bias (Schaeck et al., 2012), and hence the resulting evidence can be influenced by a confounding factor rather than the true underlying relationship between the variables of interest (Ho et al., 2007; Dunlap et al., 1996).”
Improving propensity score weighting using machine learning (2010, background): “Although some researchers have recommended checking balance on interactions [36, 37], unfortunately it is rarely done in practice.”
Analyst Initiations of Coverage and Stock Return Synchronicity (2012, positive): “7 Irvine (2003) validates his sample of initiations by searching the Dow Jones News Service (DJNS) for press coverage of the initiations.”
Foreign Aid Shocks as a Cause of Violent Armed Conflict (2011, positive): “Our matching is not exact due to the continuous nature of some of our variables and large preexisting differences between treatment and control groups, so we follow the suggestion of Ho et al. (2007) and estimate our original model22 on this matched subsample, similar to the second stage of a select”
The Causal Effect of Campus Residency on College Student Retention (2011, methodology): “Moreover, by preprocessing the data through matching, the regression estimates become less dependent on modeling choice because it reduces the link between the treatment variable (living on campus) and the control variables (Ho et al., 2007).”
When Less Is More: Effects of Grade Skipping on Adult STEM Productivity Among Mathematically Precocious Adolescents (2013, methodology): “…than .25, which has been suggested as the maximum allowed difference to grant the equivalence of randomization to a quasi-experimental or observational design (further adjustments were made using logistic regression, described below; Cochran, 1968; Ho et al., 2007; What Works Clearinghouse, 2009).”