Data Fabrication, Falsification, and Image Manipulation
Core summary
Research misconduct (FFP) encompasses fabrication (making up data), falsification (manipulating data, equipment, or processes), and plagiarism. Image manipulation — splicing, duplicating, or enhancing images beyond acceptable limits — is an increasingly detected form of falsification. The consequences include retraction, career termination, legal penalties, and harm to patients who relied on fraudulent evidence.
Detailed explanation
Detailed explanation
Fabrication is the most egregious form: inventing data points, experiments, or entire studies that were never conducted. Falsification is more subtle: selectively omitting inconvenient data points, manipulating statistical analyses to get desired p-values, altering images, or misrepresenting methods. Image manipulation has become a major focus because digital tools make alterations easy: duplicating Western blot bands, splicing gel images, adjusting contrast to hide or reveal bands, or reusing the same microscopy image across different experiments. Journals now use forensic software to detect image manipulation, and some employ dedicated image integrity analysts. Red flags include: results that seem 'too clean' (no outliers, perfect dose-response curves), unusual patterns in raw data (repeated decimal sequences, implausible distributions), images with visible splicing artifacts, and inability to provide raw data upon request. The Office of Research Integrity (ORI) in the US investigates misconduct in federally funded research. Findings of misconduct can result in: retraction of publications, debarment from federal funding (typically 3-10 years), termination of employment, and in rare cases, criminal prosecution. Beyond individual consequences, fabricated evidence that enters clinical practice can harm patients — drugs approved based on false data may be ineffective or dangerous.
Clinical example
An anesthesiology researcher fabricated data in dozens of studies over two decades. When the fraud was discovered, over 180 papers were retracted — the largest retraction case in medical history at that time. Clinical guidelines that had cited these studies had to be reassessed, potentially affecting patient care protocols worldwide.
Research example
Studies estimate that 2-3% of scientists admit to fabricating or falsifying data at least once, and up to 14% report knowing colleagues who have done so. The true prevalence is likely higher, as misconduct often goes undetected for years — if it is detected at all.
Knowledge check
Q1. What is the difference between fabrication and falsification?
Q2. Which of the following is a red flag for data integrity issues?
Q3. How does fabricated data in medical research harm patients?