Estimating the Effect of the File Drawer Problem in Meta-Analysis

1997 ◽  
Vol 85 (2) ◽  
pp. 719-722 ◽  
Author(s):  
M. T. Bradley ◽  
R. D. Gupta
Keyword(s):  
Effect Size ◽  
Meta Analysis ◽  
True Effect Size ◽  
True Effect ◽  
File Drawer ◽  
File Drawer Problem

Although meta-analysis appears to be a useful technique to verify the existence of an effect and to summarize large bodies of literature, there are problems associated with its use and interpretation. Amongst difficulties is the “file drawer problem.” With this problem it is assumed that a certain percentage of studies are not published or are not available to be included in any given meta-analysis. We present a cautionary table to quantify the magnitude of this problem. The table shows that distortions exaggerating the effect size are substantial and that the exaggerations of effects are strongest when the true effect size approaches zero. A meta-analysis could be very misleading were the true effect size close to zero.

scholarly journals The Effect of Publication Bias on the Assessment of Heterogeneity

2017 ◽  
Author(s):  
Hilde Augusteijn ◽  
Robbie Cornelis Maria van Aert ◽  
Marcel A. L. M. van Assen
Keyword(s):  
Publication Bias ◽  
Effect Size ◽  
Type I Error ◽  
Meta Analysis ◽  
Error Rates ◽  
Population Heterogeneity ◽  
Type I ◽  
Monte Carlo Simulation Study ◽  
True Effect Size ◽  
True Effect

One of the main goals of meta-analysis is to test and estimate the heterogeneity of effect size. We examined the effect of publication bias on the Q-test and assessments of heterogeneity, as a function of true heterogeneity, publication bias, true effect size, number of studies, and variation of sample sizes. The expected values of heterogeneity measures H2 and I2 were analytically derived, and the power and the type I error rate of the Q-test were examined in a Monte-Carlo simulation study. Our results show that the effect of publication bias on the Q-test and assessment of heterogeneity is large, complex, and non-linear. Publication bias can both dramatically decrease and increase heterogeneity. Extreme homogeneity can occur even when the population heterogeneity is large. Particularly if the number of studies is large and population effect size is small, publication bias can cause both extreme type I error rates and power of the Q-test close to 0 or 1. We therefore conclude that the Q-test of homogeneity and heterogeneity measures H2 and I2 are generally not valid in assessing and testing heterogeneity when publication bias is present, especially when the true effect size is small and the number of studies is large. We introduce a web application, Q-sense, which can be used to assess the sensitivity of the Q-test to publication bias, and we apply it to two published meta-analysis. Meta-analytic methods should be enhanced in order to be able to deal with publication bias in their assessment and tests of heterogeneity.

scholarly journals Bayesian Hypothesis Testing and Estimation under the Marginalized Random-Effects Meta-Analysis Model

2020 ◽  
Author(s):  
Robbie Cornelis Maria van Aert ◽  
Joris Mulder
Keyword(s):  
Hypothesis Testing ◽  
Random Effects ◽  
Effect Size ◽  
Meta Analysis ◽  
Bayes Factors ◽  
Estimation Methods ◽  
True Effect Size ◽  
Bayesian Hypothesis Testing ◽  
True Effect ◽  
Meta Analyses

Meta-analysis methods are used to synthesize results of multiple studies on the same topic. The most frequently used statistical model in meta-analysis is the random-effects model containing parameters for the average effect, between-study variance in primary study's true effect size, and random effects for the study specific effects. We propose Bayesian hypothesis testing and estimation methods using the marginalized random-effects meta-analysis (MAREMA) model where the study specific true effects are regarded as nuisance parameters which are integrated out of the model. A flat prior distribution is placed on the overall effect size in case of estimation and a proper unit information prior for the overall effect size is proposed in case of hypothesis testing. For the between-study variance in true effect size, a proper uniform prior is placed on the proportion of total variance that can be attributed to between-study variability. Bayes factors are used for hypothesis testing that allow testing point and one-sided hypotheses. The proposed methodology has several attractive properties. First, the proposed MAREMA model encompasses models with a zero, negative, and positive between-study variance, which enables testing a zero between-study variance as it is not a boundary problem. Second, the methodology is suitable for default Bayesian meta-analyses as it requires no prior information about the unknown parameters. Third, the methodology can even be used in the extreme case when only two studies are available, because Bayes factors are not based on large sample theory. We illustrate the developed methods by applying it to two meta-analyses and introduce easy-to-use software in the R package BFpack to compute the proposed Bayes factors.

scholarly journals Dissertation R.C.M. van Aert

2018 ◽  
Author(s):  
Robbie Cornelis Maria van Aert
Keyword(s):  
Publication Bias ◽  
Effect Size ◽  
Meta Analysis ◽  
Original Study ◽  
The Other ◽  
Effect Sizes ◽  
True Effect Size ◽  
True Effect ◽  
Meta Analyses ◽  
Q Statistic

More and more scientific research gets published nowadays, asking for statistical methods that enable researchers to get an overview of the literature in a particular research field. For that purpose, meta-analysis methods were developed that can be used for statistically combining the effect sizes from independent primary studies on the same topic. My dissertation focuses on two issues that are crucial when conducting a meta-analysis: publication bias and heterogeneity in primary studies’ true effect sizes. Accurate estimation of both the meta-analytic effect size as well as the between-study variance in true effect size is crucial since the results of meta-analyses are often used for policy making. Publication bias distorts the results of a meta-analysis since it refers to situations where publication of a primary study depends on its results. We developed new meta-analysis methods, p-uniform and p-uniform*, which estimate effect sizes corrected for publication bias and also test for publication bias. Although the methods perform well in many conditions, these and the other existing methods are shown not to perform well when researchers use questionable research practices. Additionally, when publication bias is absent or limited, traditional methods that do not correct for publication bias outperform p¬-uniform and p-uniform*. Surprisingly, we found no strong evidence for the presence of publication bias in our pre-registered study on the presence of publication bias in a large-scale data set consisting of 83 meta-analyses and 499 systematic reviews published in the fields of psychology and medicine. We also developed two methods for meta-analyzing a statistically significant published original study and a replication of that study, which reflects a situation often encountered by researchers. One method is a frequentist whereas the other method is a Bayesian statistical method. Both methods are shown to perform better than traditional meta-analytic methods that do not take the statistical significance of the original study into account. Analytical studies of both methods also show that sometimes the original study is better discarded for optimal estimation of the true effect size. Finally, we developed a program for determining the required sample size in a replication analogous to power analysis in null hypothesis testing. Computing the required sample size with the method revealed that large sample sizes (approximately 650 participants) are required to be able to distinguish a zero from a small true effect.Finally, in the last two chapters we derived a new multi-step estimator for the between-study variance in primary studies’ true effect sizes, and examined the statistical properties of two methods (Q-profile and generalized Q-statistic method) to compute the confidence interval of the between-study variance in true effect size. We proved that the multi-step estimator converges to the Paule-Mandel estimator which is nowadays one of the recommended methods to estimate the between-study variance in true effect sizes. Two Monte-Carlo simulation studies showed that the coverage probabilities of Q-profile and generalized Q-statistic method can be substantially below the nominal coverage rate if the assumptions underlying the random-effects meta-analysis model were violated.

Effect Size Estimation From t-Statistics in the Presence of Publication Bias

2018 ◽  
Vol 226 (1) ◽  
pp. 56-80 ◽  
Author(s):  
Rolf Ulrich ◽  
Jeff Miller ◽  
Edgar Erdfelder
Keyword(s):  
Publication Bias ◽  
Effect Size ◽  
Effect Sizes ◽  
Size Estimation ◽  
True Effect Size ◽  
True Effect ◽  
Effect Size Estimation ◽  
T Values ◽  
File Drawer

Abstract. Publication bias hampers the estimation of true effect sizes. Specifically, effect sizes are systematically overestimated when studies report only significant results. In this paper we show how this overestimation depends on the true effect size and on the sample size. Furthermore, we review and follow up methods originally suggested by Hedges (1984) , Iyengar and Greenhouse (1988) , and Rust, Lehmann, and Farley (1990) allowing the estimation of the true effect size from published test statistics (e.g., from the t-values of reported significant results). Moreover, we adapted these methods allowing meta-analysts to estimate the percentage of researchers who consign undesired results in a research domain to the file drawer. We also apply the same logic to the case when significant results tend to be underreported. We demonstrate the application of these procedures for conventional one-sample and two-sample t-tests. Finally, we provide R and MATLAB versions of a computer program to estimate the true unbiased effect size and the prevalence of publication bias in the literature.

How to Detect Publication Bias in Psychological Research

2019 ◽  
Vol 227 (4) ◽  
pp. 261-279 ◽  
Author(s):  
Frank Renkewitz ◽  
Melanie Keiner
Keyword(s):  
Publication Bias ◽  
Effect Size ◽  
Statistical Power ◽  
Type I Error ◽  
Psychological Research ◽  
Type I ◽  
True Effect Size ◽  
Questionable Research Practices ◽  
True Effect ◽  
Meta Analyses

Abstract. Publication biases and questionable research practices are assumed to be two of the main causes of low replication rates. Both of these problems lead to severely inflated effect size estimates in meta-analyses. Methodologists have proposed a number of statistical tools to detect such bias in meta-analytic results. We present an evaluation of the performance of six of these tools. To assess the Type I error rate and the statistical power of these methods, we simulated a large variety of literatures that differed with regard to true effect size, heterogeneity, number of available primary studies, and sample sizes of these primary studies; furthermore, simulated studies were subjected to different degrees of publication bias. Our results show that across all simulated conditions, no method consistently outperformed the others. Additionally, all methods performed poorly when true effect sizes were heterogeneous or primary studies had a small chance of being published, irrespective of their results. This suggests that in many actual meta-analyses in psychology, bias will remain undiscovered no matter which detection method is used.

Abstract 3789: Nessun Dorma : Have the Risks of Rosiglitazone been Exaggerated?

Circulation ◽  
2007 ◽  
Vol 116 (suppl_16) ◽  
Author(s):  
George A Diamond ◽  
Sanjay Kaul
Keyword(s):  
Confidence Intervals ◽  
Effect Size ◽  
Fixed Effects ◽  
Meta Analysis ◽  
Cardiovascular Death ◽  
Sensitivity Analyses ◽  
Odds Ratios ◽  
True Effect Size ◽  
Index Study ◽  
The Stability

Background A highly publicized meta-analysis of 42 clinical trials comprising 27,844 diabetics ignited a firestorm of controversy by charging that treatment with rosiglitazone was associated with a “…worrisome…” 43% greater risk of myocardial infarction ( p =0.03) and a 64% greater risk of cardiovascular death ( p =0.06). Objective The investigators excluded 4 trials from the infarction analysis and 19 trials from the mortality analysis in which no events were observed. We sought to determine if these exclusions biased the results. Methods We compared the index study to a Bayesian meta-analysis of the entire 42 trials (using odds ratio as the measure of effect size) and to fixed-effects and random-effects analyses with and without a continuity correction that adjusts for values of zero. Results The odds ratios and confidence intervals for the analyses are summarized in the Table . Odds ratios for infarction ranged from 1.43 to 1.22 and for death from 1.64 to 1.13. Corrected models resulted in substantially smaller odds ratios and narrower confidence intervals than did uncorrected models. Although corrected risks remain elevated, none are statistically significant (*p<0.05). Conclusions Given the fragility of the effect sizes and confidence intervals, the charge that roziglitazone increases the risk of adverse events is not supported by these additional analyses. The exaggerated values observed in the index study are likely the result of excluding the zero-event trials from analysis. Continuity adjustments mitigate this error and provide more consistent and reliable assessments of true effect size. Transparent sensitivity analyses should therefore be performed over a realistic range of the operative assumptions to verify the stability of such assessments especially when outcome events are rare. Given the relatively wide confidence intervals, additional data will be required to adjudicate these inconclusive results.

scholarly journals Coordinated Data Analysis: Knowledge Accumulation in Lifespan Developmental Psychology

2021 ◽  
Author(s):  
Eileen Kranz Graham ◽  
Emily C Willroth ◽  
Sara J Weston ◽  
Graciela Muniz-Terrera ◽  
Sean Clouston ◽  
...  
Keyword(s):  
Developmental Psychology ◽  
Meta Analysis ◽  
Research Question ◽  
Data Networks ◽  
Aging Research ◽  
Require Time ◽  
True Effect Size ◽  
Multiple Datasets ◽  
Development And Aging ◽  
File Drawer

Coordinated analysis is a powerful form of integrative analysis, and is well suited in its capacity to promote cumulative scientific knowledge, particularly in subfields of psychology that focus on the processes of lifespan development and aging. Coordinated analysis uses raw data from individual studies to create similar hypothesis tests for a given research question across multiple datasets, thereby making it less vulnerable to common criticisms of meta-analysis such as file drawer effects or publication bias. Coordinated analysis can sometimes use random effects meta-analysis to summarize results, which does not assume a single true effect size for a given statistical test. By fitting parallel models in separate datasets, coordinated analysis preserves the heterogeneity among studies, and provides a window into the generalizability and external validity of a set of results. The current paper achieves three goals: First, it describes the phases of a coordinated analysis so that interested researchers can more easily adopt these methods in their labs. Second, it discusses the importance of coordinated analysis within the context of the credibility revolution in psychology. Third, it encourages the use of existing data networks and repositories for conducting coordinated analysis, in order to enhance accessibility and inclusivity. Subfields of research that require time- or resource- intensive data collection, such as longitudinal aging research, would benefit by adopting these methods.

Bias in emerging biomarkers for bipolar disorder

2016 ◽  
Vol 46 (11) ◽  
pp. 2287-2297 ◽  
Author(s):  
A. F. Carvalho ◽  
C. A. Köhler ◽  
B. S. Fernandes ◽  
J. Quevedo ◽  
K. W. Miskowiak ◽  
...  
Keyword(s):  
Bipolar Disorder ◽  
Effect Size ◽  
Comprehensive Evaluation ◽  
Meta Analysis ◽  
Selective Reporting ◽  
True Effect Size ◽  
Genetic Biomarkers ◽  
Morning Cortisol ◽  
Meta Analyses ◽  
Positive Results

BackgroundTo date no comprehensive evaluation has appraised the likelihood of bias or the strength of the evidence of peripheral biomarkers for bipolar disorder (BD). Here we performed an umbrella review of meta-analyses of peripheral non-genetic biomarkers for BD.MethodThe Pubmed/Medline, EMBASE and PsycInfo electronic databases were searched up to May 2015. Two independent authors conducted searches, examined references for eligibility, and extracted data. Meta-analyses in any language examining peripheral non-genetic biomarkers in participants with BD (across different mood states) compared to unaffected controls were included.ResultsSix references, which examined 13 biomarkers across 20 meta-analyses (5474 BD cases and 4823 healthy controls) met inclusion criteria. Evidence for excess of significance bias (i.e. bias favoring publication of ‘positive’ nominally significant results) was observed in 11 meta-analyses. Heterogeneity was high for (I2 ⩾ 50%) 16 meta-analyses. Only two biomarkers met criteria for suggestive evidence namely the soluble IL-2 receptor and morning cortisol. The median power of included studies, using the effect size of the largest dataset as the plausible true effect size of each meta-analysis, was 15.3%.ConclusionsOur findings suggest that there is an excess of statistically significant results in the literature of peripheral biomarkers for BD. Selective publication of ‘positive’ results and selective reporting of outcomes are possible mechanisms.

scholarly journals Low statistical power in biomedical science: a review of three human research domains

2017 ◽  
Vol 4 (2) ◽  
pp. 160254 ◽  
Author(s):  
Estelle Dumas-Mallet ◽  
Katherine S. Button ◽  
Thomas Boraud ◽  
Francois Gonon ◽  
Marcus R. Munafò
Keyword(s):  
Effect Size ◽  
Statistical Power ◽  
Meta Analysis ◽  
Statistical Significance ◽  
Average Power ◽  
Biomedical Science ◽  
Significant Finding ◽  
Biomedical Sciences ◽  
True Effect Size ◽  
Meta Analyses

Studies with low statistical power increase the likelihood that a statistically significant finding represents a false positive result. We conducted a review of meta-analyses of studies investigating the association of biological, environmental or cognitive parameters with neurological, psychiatric and somatic diseases, excluding treatment studies, in order to estimate the average statistical power across these domains. Taking the effect size indicated by a meta-analysis as the best estimate of the likely true effect size, and assuming a threshold for declaring statistical significance of 5%, we found that approximately 50% of studies have statistical power in the 0–10% or 11–20% range, well below the minimum of 80% that is often considered conventional. Studies with low statistical power appear to be common in the biomedical sciences, at least in the specific subject areas captured by our search strategy. However, we also observe evidence that this depends in part on research methodology, with candidate gene studies showing very low average power and studies using cognitive/behavioural measures showing high average power. This warrants further investigation.

Sign in / Sign up

Export Citation Format

Share Document