Publication selection in health policy research: The winner's curse hypothesis

Costa‐Font, Joan; McGuire, Alistair; Stanley, T. D.

doi:10.1016/j.healthpol.2012.10.015

Cited by 35 publications

(34 citation statements)

References 48 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Elsewhere,β 1 has been called the 'precision-effect estimate with SE' (PEESE) (Stanley and Doucouliagos, 2007;Havranek, 2010;Costa-Font et al, 2011;Doucouliagos et al, 2012;Stanley and Doucouliagos, 2012;Costa-Font et al, 2013). This MRA correction for publication bias was first proposed in Stanley (2006), refined and applied in Stanley and Doucouliagos (2007) Res.…”

Section: Egger Regression and The Precision-effect Testmentioning

confidence: 99%

Meta-regression approximations to reduce publication selection bias

2013

View full text Add to dashboard Cite

Publication selection bias is a serious challenge to the integrity of all empirical sciences. We derive meta-regression approximations to reduce this bias. Our approach employs Taylor polynomial approximations to the conditional mean of a truncated distribution. A quadratic approximation without a linear term, precision-effect estimate with standard error (PEESE), is shown to have the smallest bias and mean squared error in most cases and to outperform conventional meta-analysis estimators, often by a great deal. Monte Carlo simulations also demonstrate how a new hybrid estimator that conditionally combines PEESE and the Egger regression intercept can provide a practical solution to publication selection bias. PEESE is easily expanded to accommodate systematic heterogeneity along with complex and differential publication selection bias that is related to moderator variables. By providing an intuitive reason for these approximations, we can also explain why the Egger regression works so well and when it does not. These meta-regression methods are applied to several policy-relevant areas of research including antidepressant effectiveness, the value of a statistical life, the minimum wage, and nicotine replacement therapy.

show abstract

Section: Egger Regression and The Precision-effect Testmentioning

confidence: 99%

Meta-regression approximations to reduce publication selection bias

2013

View full text Add to dashboard Cite

show abstract

“…Values of random heterogeneity,

σ_{h}^{2}

, were selected to encompass what is found in actual meta‐analyses, as measured by I 2 (Higgins and Thompson, ). For example, I 2 is 90% among US minimum wage elasticities (Doucouliagos and Stanley, ), 87% for efficiency wage elasticities (Krassoi Peach and Stanley, ), 93% among estimates of the value of statistical life (Doucouliagos, Stanley and Giles, ), 97% among the partial correlations of CEO pay and corporate performance (Doucouliagos, et al , ), 99.2% among the income elasticities of health care (Costa‐Font et al , ), and 84% among the partial correlation coefficients of UK's minimum wage increases and employment (De Linde Leonard et al , ). Needless to say, smaller values of I 2 can also be found throughout the social and medical sciences.…”

Section: Simulationsmentioning

confidence: 99%

“…D. STANLEY AND H. DOUCOULIAGOSMeth. 2017, 8 19-42 correlations of CEO pay and corporate performance, 99.2% among the income elasticities of health care(Costa-Font et al, 2013), and 84% among the partial correlation coefficients of UK's minimum wage increases and employment (De Linde Leonard et al…”

mentioning

confidence: 99%

Neither fixed nor random: weighted least squares meta‐regression

Stanley

Doucouliagos

2016

Research Synthesis Methods

190

156

View full text Add to dashboard Cite

Our study revisits and challenges two core conventional meta-regression estimators: the prevalent use of 'mixed-effects' or random-effects meta-regression analysis and the correction of standard errors that defines fixed-effects meta-regression analysis (FE-MRA). We show how and explain why an unrestricted weighted least squares MRA (WLS-MRA) estimator is superior to conventional random-effects (or mixed-effects) meta-regression when there is publication (or small-sample) bias that is as good as FE-MRA in all cases and better than fixed effects in most practical applications. Simulations and statistical theory show that WLS-MRA provides satisfactory estimates of meta-regression coefficients that are practically equivalent to mixed effects or random effects when there is no publication bias. When there is publication selection bias, WLS-MRA always has smaller bias than mixed effects or random effects. In practical applications, an unrestricted WLS meta-regression is likely to give practically equivalent or superior estimates to fixed-effects, random-effects, and mixed-effects meta-regression approaches. However, random-effects meta-regression remains viable and perhaps somewhat preferable if selection for statistical significance (publication bias) can be ruled out and when random, additive normal heterogeneity is known to directly affect the 'true' regression coefficient. Copyright © 2016 John Wiley & Sons, Ltd.

show abstract

“…However, we do not wish to overemphasize this because in 26 the government consumption sample, we find that studies published in higher-ranked journals tend to report more adverse effects of government size on growth. This is an indication of the 'Winner's curse' -whereby journals with good reputation capitalize on their reputation and publish 'more selected' findings (see Costa-Font et al, 2013;Ugur, 2014).…”

Section: Discussionmentioning

confidence: 99%

Does Government Size Affect Per-Capita Income Growth? A Hierarchical Meta-Regression Analysis

Awaworyi Churchill,

Ugur,

Yew

2015

SSRN Journal

View full text Add to dashboard Cite

Abstract:We conduct a hierarchical meta-regression analysis to review 87 empirical studies that report 769 estimates for the effects of government size on economic growth. We follow best-practice recommendations for meta-analysis of economics research, and address issues of publication selection bias and heterogeneity. When size is measured as the ratio of total government expenditures to GDP, the partial correlation between government size and per-capita GDP growth is negative in developed countries, but insignificant in developing countries. When size is measured as the ratio of consumption expenditures to GDP, the partial correlation is negative in both developed and developing countries, but the effect in developing countries is less adverse. We also report that government size is associated with less adverse effects when primary studies control for endogeneity and are published in journals and more recently, but it is associated with more adverse effects when primary studies use cross-section data. Our findings indicate that the relationship between government size and per-capita GDP growth is context-specific and likely to be biased due to endogeneity between the level of per-capita income and government expenditures. JEL code: O40, H50, C1Keywords: Economic growth; Government size; Government expenditure; Government consumption; Meta-analysis; Evidence synthesis Acknowledgements: We would like to thank two anonymous reviewers of the Program Committee of MAER-Net Colloquium 2015 in Prague; and the participants in the conference for helpful comments.

show abstract

Publication selection in health policy research: The winner's curse hypothesis

Cited by 35 publications

References 48 publications

Meta-regression approximations to reduce publication selection bias

Meta-regression approximations to reduce publication selection bias

Neither fixed nor random: weighted least squares meta‐regression

Does Government Size Affect Per-Capita Income Growth? A Hierarchical Meta-Regression Analysis

Contact Info

Product

Resources

About