The Heckman Correction for Sample Selection and Its Critique - A Short Survey

By Patrick A. Puhani. 2000

1,264 citations recorded by the source corpus.

Subfields: Economics and Econometrics, Marketing

View Paper: https://www.semanticscholar.org/paper/21ef9869b814cbdc91557663b18735e8968120e7

How researchers see this paper

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

In Economics and Econometrics

In Marketing

Citations in the corpus

Measuring teaching quality in higher education: assessing selection bias in course evaluations (2017, background): “However, as is well known, this is empirically problematic and therefore not recommended (e.g. see Keane 1992; Meng and Schmidt 1985; Puhani 2000).”
Foreign Direct Investment and R&D: Substitutes or Complements—A Case of Indian Manufacturing after 1991 Reforms (2011, background): “In reality, such variables are hard to find if the process involved in selection and response are identical (Vella 1998; Puhani 2000).”
R&D and Subsidies at the Firm Level: An Application of Parametric and Semi-Parametric Two-Step Selection Models (2008, methodology): “11See Puhani (2000) for an analysis of the usefullness of Heckman’s (1976, 1979) estimator.”
Understanding Neighbourhood Effects: Selection Bias and Residential Mobility (2010, methodology): “Although the two-step model has some disadvantages and has been criticized in the econometric literature (e.g. Puhani, 2000), it has the enormous benefit of including real measures of selective mobility and neighbourhood sorting.”
SME Policy, Financial Structure and Firm Growth: Evidence From Japan (2006, methodology): “In particular, Puhani (2000) showed that the subsample ordinary least squares method (OLS) performs better than the sample selection model if collinearity problems between two regression equations are present.”
The microeconometric estimation of treatment effects—An overview (2006, background): “…are assumed to be heterogeneous.15 The Heckman selection estimator is not without critique, which rests mainly on the following point (see e.g. Puhani (2000)): If there are no exclusion restrictions, the models are identified only by assumptions about functional form and error distributions.”
Deep generative models for reject inference in credit scoring (2019, background): “This happens either when the correlation between the error terms in the selection and outcome equations is high or the data has high degree of censoring [43].”
“Can’t Stop, Won’t Stop”: Self-Control, Risky Lifestyles, and Repeat Victimization (2012, methodology): “As Bushway and colleagues (2007: 159) point out, ‘‘when the error assumptions are met the FIML will always be more efficient than the Heckman twostep, a fact which has been demonstrated in numerous simulation studies’’ (see also Leung and Yu 1996; Puhani 2000).”
Reject inference in application scorecards: evidence from France (2016, methodology): “In the sense of Rubin (1976) and Little and Rubin (2002), missing data can be classified in one of three groups: missing completely at random (MCAR), missing at random (MAR), and missing not at random (MNAR).”
The use of propensity score matching in the evaluation of active labour market policies (2002, background): “4 See, for example, Goldberger (1983) and Puhani (2000) 5 In practice, participants may be matched to multiple non-participants.”