The File Draw Problem

The file drawer problem (a primary cause of publication bias) occurs when researchers only publish studies with positive, statistically significant results, while studies showing null, neutral, or negative findings are filed away and never seen by the scientific community.

Coined by psychologist Robert Rosenthal in 1979, the metaphor describes an extreme scenario where a few published papers claim an effect exists, while dozens of contradictory studies sit quietly in researchers' file drawers.

Why It Happens

Editorial Bias: Academic journals favor novel, ground-breaking findings. A paper stating "Drug X significantly reduces headaches" is far more likely to get accepted than "Drug X shows no difference compared to placebo."
Career Incentives: Grants, tenure, and academic prestige often depend on high-impact publications, incentivizing researchers not to spend time writing up null results.
Researcher Skepticism: When a study finds no effect, researchers often assume they made an experimental error, had poor sample quality, or used the wrong instruments, choosing not to publish.

Major Consequences

Inflated Effect Sizes: Meta-analyses that synthesize all available literature will overestimate how well a treatment, drug, or intervention actually works.
Wasted Resources: Other scientists repeatedly duplicate failed experiments because they don't know the hypothesis was already tested and disproven.
The Replication Crisis: When subsequent researchers try to reproduce published landmark studies, they often fail because the original published study was merely a statistical fluke (a False Positive, or Type I error) among dozens of hidden negative trials.

Modern Solutions

Study Preregistration & Registered Reports: Platforms like OSF and journals that accept protocols before data collection commit to publishing the results regardless of whether they turn out positive or null.
Funnel Plots: Meta-analysts plot study sample sizes against effect sizes. If small studies with null effects are missing (creating an asymmetric "gap" in the funnel shape), publication bias is likely present.
Clinical Trial Registries: Mandates (such as ClinicalTrials.gov) require researchers to register clinical experiments in advance, making unpublished trials traceable.