A meta-analysis of 50 published trials on a new drug shows a large positive effect. The funnel plot displays a marked gap in the lower-left quadrant — small studies with small or negative effects are conspicuously absent. What does this pattern most strongly suggest?
AThe meta-analysis is highly reliable because 50 studies is a large sample
BPublication bias has likely inflated the pooled effect estimate
CThe drug is ineffective, as shown by the missing negative studies
DFunnel plot asymmetry proves that the small studies were methodologically flawed
The gap in the lower-left quadrant — where small studies with null or negative results should appear — is the signature pattern of publication bias. Negative and null studies are systematically less likely to be published (the file drawer problem), so the published literature is a biased sample. This inflates the meta-analytic estimate. Option A is the classic misconception: more published studies does not reduce this bias — pooling a larger biased sample just yields a more precise but still inflated answer. Option C inverts the logic; missing studies suggest they weren't published, not that they show a specific result. Option D is wrong because asymmetry reflects publication bias, not methodological flaw.
Question 2 Multiple Choice
What does Egger's regression test for in the context of publication bias detection?
AWhether individual studies used random allocation
BWhether the pooled effect estimate is statistically significant
CWhether there is statistically detectable asymmetry in the funnel plot
DWhether all studies used the same outcome definition
Egger's regression regresses each study's standardized effect estimate on its standard error. A non-zero intercept indicates systematic asymmetry in the funnel plot — the kind expected when small studies disproportionately show larger effects, as occurs with publication bias. It quantifies what the eye detects visually in the funnel plot. It does not assess randomization, statistical significance of the pooled estimate, or outcome consistency.
Question 3 True / False
The trim-and-fill method both removes asymmetric outlier studies and imputes hypothetical missing studies, then re-estimates the pooled effect to account for likely unpublished evidence.
TTrue
FFalse
Answer: True
This is an accurate description of the trim-and-fill method. In the 'trim' step, studies on the over-represented side of the funnel are iteratively removed to estimate the true center. In the 'fill' step, hypothetical mirror-image studies are imputed on the under-represented side. The adjusted pooled estimate reflects what the meta-analysis might look like if missing studies had been published. A substantially different adjusted estimate is a red flag that the original pooled effect was inflated.
Question 4 True / False
Adding more published studies to a meta-analysis usually reduces the distortion caused by publication bias, because larger samples yield more accurate estimates.
TTrue
FFalse
Answer: False
This is the most dangerous misconception about publication bias. Adding more published studies increases precision — but if those studies are themselves subject to the same publication bias (positive results favored), pooling them simply yields a more precise estimate of the inflated value. Publication bias is systematic, not random, so increasing sample size does not average it out. The problem is not noise but selection: the literature is a biased sample of all studies conducted, and more observations from a biased sample compound rather than correct the bias.
Question 5 Short Answer
Why does publication bias cause meta-analytic effect estimates to be inflated rather than simply imprecise, and why does this make it a more serious threat than random sampling error?
Think about your answer, then reveal below.
Model answer: Publication bias is a systematic distortion, not random error. Because studies with null or negative results are less likely to be published, the available literature overrepresents positive findings. A meta-analysis pools these overrepresented positive studies, yielding an estimate that is biased upward. Unlike random error — which averages out with more data — systematic bias grows as more biased studies are pooled: the estimate becomes more precise but remains systematically wrong. This is why even large meta-analyses cannot self-correct for publication bias without external evidence of missing studies.
The key distinction is systematic vs. random error. Random error cancels out across studies; systematic error compounds. Publication bias introduces a directional filter that means the published record is not a random sample of the evidence — it is a positively selected sample. Any synthesis of that record inherits and amplifies the selection effect.