## Estimating and Adjusting for Publication Bias Using Data Augmentation in Bayesian Meta-Analysis (1995)

Citations: | 4 - 1 self |

### BibTeX

@TECHREPORT{Givens95estimatingand,

author = {Geof Givens and D. D. Smith and R. L. Tweedie},

title = {Estimating and Adjusting for Publication Bias Using Data Augmentation in Bayesian Meta-Analysis},

institution = {},

year = {1995}

}

### OpenURL

### Abstract

We introduce a Bayesian approach which estimates and adjusts for selection bias in a set of studies used in a meta-analysis. We use a hierarchical model for study outcome, and propose an additional model component to account for publication bias, which is the possibility that studies of interest are not equally likely to be published and hence observed studies are not a random sample. Estimation is based on the data augmentation principle and the number and outcomes of unobserved studies are simulated using Gibbs sampling methods. After examining simulation performance, we apply our techniques to a meta-analysis of 35 studies of the relationship between lung cancer and spousal exposure to environmental tobacco smoke. We find that the 95% posterior probability interval for relative risk is shifted downward after allowing for this. These results are consistent with earlier, ad hoc, approaches to this problem. Keywords and phrases: Meta-analysis, publication bias, missing studies, Markov...

### Citations

8089 | Maximum likelihood from incomplete data via the EM algorithm
- Dempster, Laird, et al.
- 1977
(Show Context)
Citation Context ..., thus creating a `complete' dataset for analysis. Data augmentation is a technique which has proven useful in a range of Bayesian and likelihood problems, including applications of the EM algorithm (=-=Dempster et al., 1977-=-) and the IP algorithm (Tanner and Wong, 1987). The premise of data augmentation is that the `observed data' Y can be thought of as a partial realization of the random variable X = (Y; Z), where a com... |

3719 |
Stochastic Relaxation, Gibbs Distributions and the Bayesian Restoration of Images
- Geman, Geman
- 1984
(Show Context)
Citation Context ...an the standard hierarchical Bayesian model, and the posterior for \Delta can no longer be derived in a simple analytical form. Instead, simulation techniques must be used, and we use Gibbs sampling (=-=Geman and Geman, 1984-=-) to obtain approximate samples from the desired posterior distribution. Gibbs sampling techniques, which have been very successful at solving a wide variety of similar problems in Bayesian estimation... |

549 |
The calculation of posterior distributions by data augmentation
- Tanner, Wong
- 1987
(Show Context)
Citation Context ...ysis. Data augmentation is a technique which has proven useful in a range of Bayesian and likelihood problems, including applications of the EM algorithm (Dempster et al., 1977) and the IP algorithm (=-=Tanner and Wong, 1987-=-). The premise of data augmentation is that the `observed data' Y can be thought of as a partial realization of the random variable X = (Y; Z), where a complete realization, X, of X is called the `com... |

362 |
Statistical methods for meta-analysis
- Hedges, Olkin
- 1985
(Show Context)
Citation Context ...ction Meta-analysis seeks to combine the analyses from individual studies into a single statistical analysis with an overall estimate and confidence interval for effect size (Cooper and Hedges, 1994; =-=Hedges and Olkin, 1985-=-). Ideally, greater statistical power can be achieved through meta-analysis than through any one individual study, since data from a greater number of subjects are used, and in recent years there has ... |

351 |
Bayesian computation via the Gibbs Sampler and related Markov Chain Monte Carlo Method
- Smith, Roberts
- 1993
(Show Context)
Citation Context ...to obtain approximate samples from the desired posterior distribution. Gibbs sampling techniques, which have been very successful at solving a wide variety of similar problems in Bayesian estimation (=-=Smith and Roberts, 1993-=-; Besag and Green, 1993), can be used to obtain a sample from a desired distribution by simulating realizations from a Markov chain whose stationary distribution is equal to the target distribution. I... |

153 | Using the SIR Algorithm to Simulate Posterior Distributions - Rubin - 1988 |

61 |
Spatial Statistics and Bayesian Computation (with
- Besag, Green
- 1993
(Show Context)
Citation Context ...ples from the desired posterior distribution. Gibbs sampling techniques, which have been very successful at solving a wide variety of similar problems in Bayesian estimation (Smith and Roberts, 1993; =-=Besag and Green, 1993-=-), can be used to obtain a sample from a desired distribution by simulating realizations from a Markov chain whose stationary distribution is equal to the target distribution. In our case, the target ... |

45 |
Summing Up: The Science of Reviewing Research
- Light, Pillemer
- 1984
(Show Context)
Citation Context ... model with a Bayesian or nonBayesian analysis described in Section 3. There have been methods developed which help one to identify visually the existence of missing studies, such as the funnel plot (=-=Light and Pillemer 1984-=-) and related graphical methods (Thompson, 1993; Mengersen et al., 1995). There have also been some recent attempts to evaluate just how many missing studies there might be, and what their outcomes ar... |

40 | The maximal smoothing principle in density estimation - Terrell - 1990 |

36 | Introduction to the theory of statistics. Third edition - Mood, Graybill, et al. - 1974 |

35 | Non-smoking wives of heavy smokers have a higher risk of lung cancer: a study from Japan .Br Med J - Hirayama |

33 |
The handbook of research synthesis. Russel Sage Foundation
- Cooper, Hedges
- 1994
(Show Context)
Citation Context ...drawer problem. 1 Introduction Meta-analysis seeks to combine the analyses from individual studies into a single statistical analysis with an overall estimate and confidence interval for effect size (=-=Cooper and Hedges, 1994-=-; Hedges and Olkin, 1985). Ideally, greater statistical power can be achieved through meta-analysis than through any one individual study, since data from a greater number of subjects are used, and in... |

32 | Comment on “The calculation of posterior distributions by data augmentation”, by M.Tanner and W.H.Wang - Rubin - 1987 |

23 | Bayesian Metaanalysis - DuMouchel - 1990 |

21 |
Modeling publication selection effects in meta-analysis
- Hedges
- 1992
(Show Context)
Citation Context ...techniques (see for example Mosteller and Chalmers, 1992; Felson, 1992; Chalmers, 1991; NRC Report, 1992; Thompson and Pocock, 1991; Mengersen et al., 1995). One of the most well documented concerns (=-=Hedges, 1992-=-; Dear and Begg 1992; Sterling et al., 1995) is the need to have available all relevant information. It is clearly crucial to attempt to collect at least all published studies; if possible, one should... |

15 |
An approach for assessing publication bias prior to performing a meta-analysis
- Dear, Begg
- 1992
(Show Context)
Citation Context ...e for example Mosteller and Chalmers, 1992; Felson, 1992; Chalmers, 1991; NRC Report, 1992; Thompson and Pocock, 1991; Mengersen et al., 1995). One of the most well documented concerns (Hedges, 1992; =-=Dear and Begg 1992-=-; Sterling et al., 1995) is the need to have available all relevant information. It is clearly crucial to attempt to collect at least all published studies; if possible, one should also search for unp... |

13 |
Selection models and the file drawer problem
- Iyengar, Greenhouse
- 1988
(Show Context)
Citation Context ...ct of unpublished or uncollected studies, but this has been rarely done so far. In this paper we provide a new approach to the formal evaluation of such a `publication bias' or `file drawer problem' (=-=Iyengar and Greenhouse, 1988-=-). Our method is based on a Bayesian hierarchical model for meta-analysis that combines the estimated effect sizes from heterogeneous individual studies after estimating and adjusting for potential pu... |

13 |
The impact of method choice on meta-analysis
- Mengersen, Tweedie, et al.
- 1995
(Show Context)
Citation Context ...ppeared which assess the benefits, drawbacks and problems of these techniques (see for example Mosteller and Chalmers, 1992; Felson, 1992; Chalmers, 1991; NRC Report, 1992; Thompson and Pocock, 1991; =-=Mengersen et al., 1995-=-). One of the most well documented concerns (Hedges, 1992; Dear and Begg 1992; Sterling et al., 1995) is the need to have available all relevant information. It is clearly crucial to attempt to collec... |

11 | Health effects of passive smoking: assessment of lung cancer in adults and respiratory disorders in children. United States EPA - REVIEW - 1990 |

11 | Environmental Tobacco Smoke and Mortality - Lee - 1992 |

9 |
Tainted truth: The manipulation of fact in
- Crossen
- 1994
(Show Context)
Citation Context ...r PhD or MS dissertations; or studies are suppressed by those who do not wish to have results appear that are against their own vested interests, political beliefs, or funding source's interests (see =-=Crossen, 1994-=-, p. 19). With these possible reasons for publication bias, it is clearly hard to ensure that all studies will be found even by diligent search procedures. Standard meta-analyses can be misleading whe... |

8 |
Publication decisions revisited: The effect of the outcome of statistical tests on the decision to publish and vice versa
- Sterling, Rosenbaum, et al.
- 1995
(Show Context)
Citation Context ...ler and Chalmers, 1992; Felson, 1992; Chalmers, 1991; NRC Report, 1992; Thompson and Pocock, 1991; Mengersen et al., 1995). One of the most well documented concerns (Hedges, 1992; Dear and Begg 1992; =-=Sterling et al., 1995-=-) is the need to have available all relevant information. It is clearly crucial to attempt to collect at least all published studies; if possible, one should also search for unpublished studies such a... |

7 | An Assessment of Publication Bias Using a Sample of Published Clinical Trials - Berlin, Begg, et al. - 1989 |

7 |
Meta-analysis: methods for combining independent studies
- Olkin
- 1992
(Show Context)
Citation Context ...cent years there has been an enormous increase in the use of meta-analysis in many medical areas in order to obtain overall evaluations of association in areas where individual studies are equivocal (=-=Olkin, 1992-=-). Geof Givens is Assistant Professor, David Smith is Graduate Research Assistant, and Richard Tweedie is Professor and Chair, all at the Department of Statistics, Colorado State University, Fort Coll... |

6 | Lung cancer in Japan: effects of nutrition and exposure to ETS - HIRAYAMA - 1984 |

6 |
Some progress and problems in meta-analysis of clinical trials
- Mosteller, Chalmers
- 1992
(Show Context)
Citation Context ...llins, CO 80523 With this increased interest in meta-analysis, a large number of discussion papers have appeared which assess the benefits, drawbacks and problems of these techniques (see for example =-=Mosteller and Chalmers, 1992-=-; Felson, 1992; Chalmers, 1991; NRC Report, 1992; Thompson and Pocock, 1991; Mengersen et al., 1995). One of the most well documented concerns (Hedges, 1992; Dear and Begg 1992; Sterling et al., 1995)... |

6 |
Probing encountered data, meta-analysis and weighted distribution methods
- Patil, Taillie
- 1989
(Show Context)
Citation Context ...hat a reasonable amount of unsuppressed data drives the model to an appropriate conclusion. Direct parametric modeling of publication probability has also been proposed (Iyengar and Greenhouse, 1988; =-=Patil and Taillie, 1989-=-). Larose and Dey (1995) survey and compare several alternative parametric models. This parametric approach seems to yield some of the same benefits as our approach, including tighter posterior confid... |

5 | Passive smoking in the workplace: Classical and bayesian meta-analyses
- Biggerstaff, Tweedie, et al.
- 1994
(Show Context)
Citation Context ...variance,soe 2 1j , is exactly correct, so p(oe 2 ) is degenerate for observed studies. For �� 2 1 , we use an exponential prior with mean 0.031, based on a meta-analysis of studies on workplace E=-=TS (Biggerstaff et al., 1994-=-). We take an improper uniform prior for Z and our initial prior for w (before being scaled by the largest) is that of 3 uniform random variates on (0:5; 1] for I 1 ; I 2 ; I 3 , and uniform on (0:3; ... |

5 |
Problems induced by meta-analysis
- Chalmers
- 1991
(Show Context)
Citation Context ... in meta-analysis, a large number of discussion papers have appeared which assess the benefits, drawbacks and problems of these techniques (see for example Mosteller and Chalmers, 1992; Felson, 1992; =-=Chalmers, 1991-=-; NRC Report, 1992; Thompson and Pocock, 1991; Mengersen et al., 1995). One of the most well documented concerns (Hedges, 1992; Dear and Begg 1992; Sterling et al., 1995) is the need to have available... |

4 |
Preexisting lung disease and lung cancer among nonsmoking women
- Alavanja, Brownson, et al.
- 1992
(Show Context)
Citation Context ...e substance is a potential carcinogen, the disease is lung cancer, and the hypothesis of concern is \Delta ? 0. Between 9% to 20% of lung cancer cases occur in non-smokers (Schneiderman et al., 1989; =-=Alavanja et al., 1992-=-). Up until the early 1980's, epidemiological studies had not reported any noticeable increase in the incidence of lung cancer among non-smokers who were exposed to ETS. This changed starting in 1981 ... |

4 | Lung cancer and exposure to ETS - TRICHOPOULOS, KALANDIDI, et al. - 1981 |

3 |
Bias in meta-analytic research
- Felson
- 1992
(Show Context)
Citation Context ...eased interest in meta-analysis, a large number of discussion papers have appeared which assess the benefits, drawbacks and problems of these techniques (see for example Mosteller and Chalmers, 1992; =-=Felson, 1992-=-; Chalmers, 1991; NRC Report, 1992; Thompson and Pocock, 1991; Mengersen et al., 1995). One of the most well documented concerns (Hedges, 1992; Dear and Begg 1992; Sterling et al., 1995) is the need t... |

3 | Lung cancer and exposure to ETS. Conclusion of Greek study - TRICHOPOULOS, KALANDIDI, et al. - 1983 |

2 | Incorporating sensitivity to estimates of heterogeneity in the random effects model - Biggerstaff, Tweedie - 1996 |

2 | Variability of coverage probabilities when applying a random effects methodology for meta-analysis. Unpublished - Larholt, Tsiatis, et al. - 1994 |

2 |
Mausner & Bahn Epidemiology--An Introductory Text, second edition, W.B
- Mausner, Kramer
- 1985
(Show Context)
Citation Context ... P r[getting disease j no exposure] : In general, studies are necessarily observational, rather than controlled experiments, and in the two most common study designs, cohort and case-control studies (=-=Mausner and Kramer, 1985-=-), subjects are categorized in a 2 \Theta 2 cross-classification table: each subject is classified as either exposed to the possible toxic agent or not exposed. Each subject is also classified based o... |

2 |
Lung cancer that is not attributable to smoking: letter to the editor
- Schneiderman, Davis, et al.
- 1989
(Show Context)
Citation Context ...smoking'). In this case, the substance is a potential carcinogen, the disease is lung cancer, and the hypothesis of concern is \Delta ? 0. Between 9% to 20% of lung cancer cases occur in non-smokers (=-=Schneiderman et al., 1989-=-; Alavanja et al., 1992). Up until the early 1980's, epidemiological studies had not reported any noticeable increase in the incidence of lung cancer among non-smokers who were exposed to ETS. This ch... |

2 |
Controversies in meta-analysis: the case of the trials of serum cholesterol reduction Statist
- Thompson
- 1993
(Show Context)
Citation Context ... in Section 3. There have been methods developed which help one to identify visually the existence of missing studies, such as the funnel plot (Light and Pillemer 1984) and related graphical methods (=-=Thompson, 1993-=-; Mengersen et al., 1995). There have also been some recent attempts to evaluate just how many missing studies there might be, and what their outcomes are (Berlin, Begg, and Louis 1989; Hedges 1992; D... |

1 |
Can meta-analyses be trusted? Lancet 338 1127--1130
- Thompson, Pocock
- 1991
(Show Context)
Citation Context ...of discussion papers have appeared which assess the benefits, drawbacks and problems of these techniques (see for example Mosteller and Chalmers, 1992; Felson, 1992; Chalmers, 1991; NRC Report, 1992; =-=Thompson and Pocock, 1991-=-; Mengersen et al., 1995). One of the most well documented concerns (Hedges, 1992; Dear and Begg 1992; Sterling et al., 1995) is the need to have available all relevant information. It is clearly cruc... |