A free tool · Risk communication, not a forecast
Two published numbers, applied to your own pace. This page does not know your practices and does not predict your future. It shows you the published base rate, lets you set your own numbers, and shows every source, so the arithmetic is yours.
Nothing here is collected or stored. Move the sliders; the math happens on your screen and nowhere else.
What the published base rate implies
Over 18 manuscripts, if each carried the published base rate of fabricated references in published biomedical papers (1 in 277, early 2026), the chance that at least one contains a fabricated reference is about:
6.3%
call it about 1 in 16
Arithmetic: 1 - (1 - 1/277)18, with n = 18. Check it yourself; that is the house style.
Read this before you read the number
The base rate is a population average across 2.5 million published biomedical papers. It is not your rate. If you verify every citation at the registry before submitting, your rate is lower than this. If AI drafted your references and no one checked them at the source, published evaluations have measured much higher rates.
This number is context for a decision, not a prediction about you, and we will not pretend otherwise. The honest use of this page is simple: decide whether the number is large enough, at your pace, to be worth a deliberate check.
And if one surfaces after publication
Among NIH-funded papers, a single misconduct retraction cost a mean of $392,582 in direct funding, before any career cost. We state that once, plainly, and leave the weighing to you.
Every number on this page, at its source
Topaz et al., “Fabricated citations: an audit across 2.5 million biomedical papers.” The Lancet, 2026. PMID 42107362. The audit also reports roughly 1 in 2,828 for 2023, so the rate is rising; this page uses the most recent figure. Check the record at PubMed.
Stern et al., “Financial costs and personal consequences of research misconduct resulting in retracted publications.” eLife, 2014. PMID 25124673. Scope: NIH-funded papers, 1992 to 2012, misconduct retractions specifically. Check the record at PubMed.
Whatever the sliders say, the useful next step is free: run the pre-submission citation checklist against your own reference list. If you would rather have the check done for you, at depth, with a person reading and releasing the record, request a review.