Results 1  10
of
79
Nonparametric estimation of average treatment effects under exogeneity: a review
 REVIEW OF ECONOMICS AND STATISTICS
, 2004
"... Recently there has been a surge in econometric work focusing on estimating average treatment effects under various sets of assumptions. One strand of this literature has developed methods for estimating average treatment effects for a binary treatment under assumptions variously described as exogen ..."
Abstract

Cited by 597 (23 self)
 Add to MetaCart
Recently there has been a surge in econometric work focusing on estimating average treatment effects under various sets of assumptions. One strand of this literature has developed methods for estimating average treatment effects for a binary treatment under assumptions variously described as exogeneity, unconfoundedness, or selection on observables. The implication of these assumptions is that systematic (for example, average or distributional) differences in outcomes between treated and control units with the same values for the covariates are attributable to the treatment. Recent analysis has considered estimation and inference for average treatment effects under weaker assumptions than typical of the earlier literature by avoiding distributional and functionalform assumptions. Various methods of semiparametric estimation have been proposed, including estimating the unknown regression functions, matching, methods using the propensity score such as weighting and blocking, and combinations of these approaches. In this paper I review the state of this
Some practical guidance for the implementation of propensity score matching
 IZA DISCUSSION PAPER
, 2005
"... ..."
Propensity Score Estimation with Boosted Regression for Evaluating Causal Effects in Observational Studies
 Psychological Methods
, 2004
"... Causal effect modeling with naturalistic rather than experimental data is challenging. In observational studies participants in different treatment conditions may also differ on pretreatment characteristics that influence outcomes. Propensity score methods can theoretically eliminate these confounds ..."
Abstract

Cited by 85 (7 self)
 Add to MetaCart
(Show Context)
Causal effect modeling with naturalistic rather than experimental data is challenging. In observational studies participants in different treatment conditions may also differ on pretreatment characteristics that influence outcomes. Propensity score methods can theoretically eliminate these confounds for all observed covariates, but accurate estimation of propensity scores is impeded by large numbers of covariates, uncertain functional forms for their associations with treatment selection, and other problems. This paper demonstrates that boosting, a modern statistical technique, can overcome many of these obstacles. We illustrate this approach with a study of adolescent probationers in substance abuse treatment programs. Propensity score weights estimated using boosting eliminate most pretreatment group differences, and substantially alter the apparent relative effects of adolescent substance abuse treatment. Experimental studies offer the most rigorous evidence with which to establish treatment efficacy, but they are not always practical or feasible. Experimental treatment evaluations can be expensive to field and may be too slow to produce answers to pressing questions. In some cases
Incentives and creativity: Evidence from the academic life sciences
 RAND Journal of Economics
, 2011
"... This paper tests the hypothesis that freedom to experiment, tolerance for early failure, long time horizons to evaluate results, and detailed feedback on performance stimulate creativity and innovation in scientific research. We do so by studying the careers and output of U.S. academic life scientis ..."
Abstract

Cited by 38 (10 self)
 Add to MetaCart
(Show Context)
This paper tests the hypothesis that freedom to experiment, tolerance for early failure, long time horizons to evaluate results, and detailed feedback on performance stimulate creativity and innovation in scientific research. We do so by studying the careers and output of U.S. academic life scientists funded through two very distinct mechanisms: investigatorinitiated R01 grants from the NIH, or appointment as an investigator of the Howard Hughes Medical Institute (HHMI), whose funding practices embody many of the elements mentioned above. Using careful adjustment for selection on observables, we find that HHMI investigators produce highimpact papers at a much higher rate than two control groups of similarlyaccomplished NIHfunded scientists. In contrast, the program does not appear to have much effect on the raw number of articles published. We also observe large effects on the probability of being elected to prestigious scientific societies or the training of students that go on to win early career prizes.
Does digital divide or provide? The impact of cell phones on grain markets in Niger
 BREAD Working Papers
, 2008
"... Abstract. Due partly to costly information, price dispersion across markets is common in developed and developing countries. Between 2001 and 2006, cell phone service was phased in throughout Niger, providing an alternative and cheaper search technology to grain traders and other market actors. We c ..."
Abstract

Cited by 24 (0 self)
 Add to MetaCart
Abstract. Due partly to costly information, price dispersion across markets is common in developed and developing countries. Between 2001 and 2006, cell phone service was phased in throughout Niger, providing an alternative and cheaper search technology to grain traders and other market actors. We construct a novel theoretical model of sequential search, in which traders engage in optimal search for the maximum sales price, net transport costs. The model predicts that cell phones will increase traders’ reservation sales prices and the number of markets over which they search, leading to a reduction in price dispersion across markets. To test the predictions of the theoretical model, we use a unique market and trader dataset from Niger that combines data on prices, transport costs, rainfall and grain production with cell phone access and trader behavior. We first exploit the quasiexperimental nature of cell phone coverage to estimate the impact of the staggered introduction of information technology on market performance. The results provide evidence that cell phones reduce grain price dispersion across markets by a minimum of 6.4 percent and reduce intraannual price variation by 10 percent. Cell phones have a greater impact on price dispersion for market pairs that are farther away, and for those with lower road quality. This effect becomes larger as a higher percentage of markets have cell phone coverage. We provide empirical evidence in support of specific mechanisms that partially explain the impact of cell phones on market performance.
A distributional approach for causal inference using propensity scores
 Journal of the American Statistical Association
, 2006
"... Drawing inferences about the effects of treatments and actions is a common challenge in economics, epidemiology, and other fields. We adopt Rubin’s potential outcomes framework for causal inference and propose two methods serving complementary purposes. One can be used to estimate average causal eff ..."
Abstract

Cited by 22 (8 self)
 Add to MetaCart
(Show Context)
Drawing inferences about the effects of treatments and actions is a common challenge in economics, epidemiology, and other fields. We adopt Rubin’s potential outcomes framework for causal inference and propose two methods serving complementary purposes. One can be used to estimate average causal effects, assuming no confounding given measured covariates. The other can be used to assess how the estimates might change under various departures from no confounding. Both methods are developed from a nonparametric likelihood perspective. The propensity score plays a central role and is estimated through a parametric model. Under the assumption of no confounding, the joint distribution of covariates and each potential outcome is estimated as a weighted empirical distribution. Expectations from the joint distribution are estimated as weighted averages or, equivalently to first order, regression estimates. The likelihood estimator is at least as efficient and the regression estimator is at least as efficient and robust as existing estimators. Regardless of the noconfounding assumption, the marginal distribution of covariates times the conditional distribution of observed outcome given each treatment assignment and covariates is estimated. For a fixed bound on unmeasured confounding, the marginal distribution of covariates times the conditional distribution of counterfactual outcome given each treatment assignment and covariates is explored to the extreme and then compared with the composite distribution corresponding to observed outcome given the same treatment assignment and covariates. We illustrate the methods by analyzing the data from an observational study on right heart catheterization.
Weighting Regressions by Propensity Scores
"... Regressions can be weighted by propensity scores in order to reduce bias. However, weighting is likely to increase random error in the estimates, and to bias the estimated standard errors downward, even when selection mechanisms are well understood. Moreover, in some cases, weighting will increase t ..."
Abstract

Cited by 20 (3 self)
 Add to MetaCart
Regressions can be weighted by propensity scores in order to reduce bias. However, weighting is likely to increase random error in the estimates, and to bias the estimated standard errors downward, even when selection mechanisms are well understood. Moreover, in some cases, weighting will increase the bias in estimated causal parameters. If investigators have a good causal model, it seems better just to fit the model without weights. If the causal model is improperly specified, there can be significant problems in retrieving the situation by weighting, although weighting may help under some circumstances.
On a Class of BiasAmplifying Variables that Endanger Effect Estimates
, 2010
"... This note deals with a class of variables that, if conditioned on, tends to amplify confounding bias in the analysis of causal effects. This class, independently discovered by Bhattacharya and Vogt (2007) and Wooldridge (2009), includes instrumental variables and variables that have greater influenc ..."
Abstract

Cited by 20 (8 self)
 Add to MetaCart
(Show Context)
This note deals with a class of variables that, if conditioned on, tends to amplify confounding bias in the analysis of causal effects. This class, independently discovered by Bhattacharya and Vogt (2007) and Wooldridge (2009), includes instrumental variables and variables that have greater influence on treatment selection than on the outcome. We offer a simple derivation and an intuitive explanation of this phenomenon and then extend the analysis to non linear models. We show that: 1. the biasamplifying potential of instrumental variables extends over to nonlinear models, though not as sweepingly as in linear models; 2. in nonlinear models, conditioning on instrumental variables may introduce new bias where none existed before; 3. in both linear and nonlinear models, instrumental variables have no effect on selectioninduced bias. 1
2008. “On the Specification of Propensity Scores: with Applications to the Analysis of Trade Policies
 Journal of Business and Economics Statistics Forthcoming
"... The use of propensity score methods for program evaluation with nonexperimental data typically requires that the propensity score be estimated, often with a model whose specification is unknown. Although theoretical results suggest that estimators using more flexible propensity score specifications ..."
Abstract

Cited by 16 (2 self)
 Add to MetaCart
The use of propensity score methods for program evaluation with nonexperimental data typically requires that the propensity score be estimated, often with a model whose specification is unknown. Although theoretical results suggest that estimators using more flexible propensity score specifications perform better, this has not filtered into applied research. Here we provide Monte Carlo evidence indicating benefits of overspecifying the propensity score that are robust across a number of different covariate structures and estimators. We illustrate these results with two applications, one assessing the environmental effects of General Agreement on Tariffs and Trade/World Trade Organization membership and the other assessing the impact of adopting the euro on bilateral trade. KEY WORDS:
Bayesian Nonparametric Modeling for Causal Inference
, 2007
"... Researchers have long struggled to identify causal effects in nonexperimental settings. Many recentlyproposed strategies assume ignorability of the treatment assignment mechanism and require fitting two models – one for the assignment mechanism and one for the response surface. We propose a strate ..."
Abstract

Cited by 16 (2 self)
 Add to MetaCart
Researchers have long struggled to identify causal effects in nonexperimental settings. Many recentlyproposed strategies assume ignorability of the treatment assignment mechanism and require fitting two models – one for the assignment mechanism and one for the response surface. We propose a strategy that instead focuses on very flexibly modeling just the response surface using a Bayesian nonparametric modeling procedure, Bayesian Additive Regression Trees (BART). BART has several advantages: it is far simpler to use than many recent competitors, requires less guesswork in model fitting, handles a large number of predictors, yields coherent uncertainty intervals, fluidly handles continuous treatment variables and missing data for the outcome variable. BART produces more efficient estimates in the nonlinear situations tested in our simulations compared to propensity score matching, propensityweighted estimators, and regression adjustment. Further, it is highly competitive in linear settings with the “correct” model, linear regression.