Propensity score matching and variations on the balancing test

Lee, Wang Sheng

doi:10.1007/s00181-011-0481-0

Cited by 158 publications

(91 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The off support observations are discarded for having poor or no matches. A standardized difference of 20 is considered to be large (Lee 2013;Rosenbaum and Rubin 1985). This means that the unmatched sample has large differences between the treated and untreated group and a simple comparison between the two may be inadequate.…”

Section: Resultsmentioning

confidence: 99%

Food insecurity and housing instability in vulnerable families

King

2016

Rev Econ Household

View full text Add to dashboard Cite

Reducing the prevalence of household food insecurity has been a long-standing objective of the federal government. Previous research has found many negative consequences of food insecurity for families and households but has not examined its relationship with housing instability. Using longitudinal data from the Fragile Families and Child Wellbeing Study, difference-in-difference models show that food insecurity is associated with housing instability. The association remains statistically significant after accounting for potential selection and unobserved heterogeneity using propensity score matching and excluding households that experienced prior housing instability from the sample. Examining potential mediating factors, I find that material hardship explains about half of this association. These findings suggest that maintaining a strong social safety net would reduce the risk that families experience material hardship and housing instability, which may also reduce the risk of homelessness.

show abstract

Section: Resultsmentioning

confidence: 99%

Food insecurity and housing instability in vulnerable families

King

2016

Rev Econ Household

View full text Add to dashboard Cite

show abstract

“…We therefore examine the quality of the matching using four tests commonly implemented in the evaluation literature (see, e.g., Smith and Todd, 2005b;Lee, 2006;and Girma and Görg, 2007;Caliendo and Kopeinig, 2008;Arnold and Javorcik, 2009). …”

Section: Appendix Assistance Effects With Continuous Outcome Variablesmentioning

confidence: 99%

“…The average reduction ranges from 76.7% to 87.0%, depending on the estimator used. Further, even though there is no formal criterion to identify a standardized bias as "large", following Rosenbaum and Rubin (1985) the usual practice is to consider biases above 20% as large (see, e.g., Smith and Todd, 2005b;Lee, 2006;and Girma and Görg, 2007). As shown in the first panel of Table A1, the standardized differences after matching do not exceed this value for all variables.…”

Section: Appendix Assistance Effects With Continuous Outcome Variablesmentioning

confidence: 99%

Entering new country and product markets: does export promotion help?

Martincus

Carballo

2010

Rev World Econ

View full text Add to dashboard Cite

Standard-Nutzungsbedingungen:Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden.Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen.Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Entering new export markets is primarily a discrete choice. Even though several empirical papers have used modeling strategies consistent with this fact, no study has examined the effects of public policies aimed at affecting this decision within this setting. In this paper we assess the impact of trade promotion activities on export outcomes using trade support and highly disaggregated export data for the whole population of exporters of a small developing country, Uruguay, over the period 2000-2007 to estimate a binary outcome model which allows for unobserved heterogeneity. We find that trade supporting activities have helped firms reach new destination countries and introduce new differentiated products. We would like to thank URUGUAY XXI which kindly provided us with export and trade promotion assistance data for Uruguayan firms. We also wish to thank Alberto Abadie, who kindly shared with us the code to implement his semi-parametric difference-indifferences estimator; and Juan Blyde, Mauricio Mesquita Moreira, and Matthew Shearer for valuable comments and suggestions. We owe gratitude to Mariana Sobral de Elia for competent editing assistance. Usual disclaimers apply. Terms of use: Documents in2

show abstract

“…Additionally, Smith and Todd (2005b) criticize the use of balancing tests per se because they lack formal criteria for determining when the balance is sufficient. In line with this argument, Lee (2013) demonstrated that balancing tests display size problems. For the DW algorithm, he found that the t-test for balance led to rejection in 23.8% of tested cases, instead of the conventional 5%.…”

Section: The Dehejia and Wahba (2002) Algorithm For Reducing Misspecimentioning

confidence: 65%

“…For the DW algorithm, he found that the t-test for balance led to rejection in 23.8% of tested cases, instead of the conventional 5%. To alleviate these high rejection rates, Lee (2013) developed a permutation version of the traditional t-test. This updated test leads to test sizes of 3.5% for the DW algorithm; thus, it is rather conservative.…”

Section: The Dehejia and Wahba (2002) Algorithm For Reducing Misspecimentioning

confidence: 99%

Erweiterung der Propensity Score Gleichung zur Vermeidung von Fehlspezifikationen? Eine Monte Carlo Simulation

Krug

2017

AStA Wirtsch Sozialstat Arch

View full text Add to dashboard Cite

Propensity score matching is a semi-parametric method of balancing covariates that estimates the causal effect of a treatment, intervention, or action on a specific outcome. Propensity scores are typically estimated using parametric models for binary outcomes, such as logistic regression. Therefore, model specification may still play an important role, even if the causal effect is estimated nonparametrically in the matched sample. Methodological research indicates that incorrect specification of the propensity score equation can lead to biased estimates. Augmenting the propensity score equation with terms that represent potential nonlinearity and nonadditivity, as proposed by Dehejia and Wahba and more recently by Imbens and Rubin, represents a means of avoiding such bias. Here, we conduct a Monte Carlo simulation and show that the misspecification bias is rather small in many situations. However, when the propensity score equation and/or the outcome equation are characterized by strong nonlinearity and nonadditivity, the misspecification bias can be severe. Augmentation is shown to reduce this bias, often substantially. The Dehejia-Wahba (2002)

show abstract

Propensity score matching and variations on the balancing test

Cited by 158 publications

References 39 publications

Food insecurity and housing instability in vulnerable families

Food insecurity and housing instability in vulnerable families

Entering new country and product markets: does export promotion help?

Erweiterung der Propensity Score Gleichung zur Vermeidung von Fehlspezifikationen? Eine Monte Carlo Simulation

Contact Info

Product

Resources

About