Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

How does correlation structure differ between real and fabricated data-sets?

Noori Akhtar-Danesh1, Mahshid Dehghan-Kooshkghazi

  • 1School of Nursing & Department of Clinical Epidemiology and Biostatistics, McMaster University, Hamilton, Canada. daneshn@mcmaster.ca

BMC Medical Research Methodology
|October 1, 2003
PubMed
Summary
This summary is machine-generated.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

A population-based study on the impact of social determinants of health on time to treatment of stage I lung cancer in Ontario, Canada.

BMC cancer·2026
Same author

Harnessing artificial intelligence for automatic feedback in a virtual anatomy study tool: A Q-methodology study.

Anatomical sciences education·2026
Same author

Healthcare Professionals' Preferences and Needs for Continuing Professional Development Activities: A Q-Methodology Study.

Teaching and learning in medicine·2026
Same author

The Impact of Social Determinants of Health on Treatment Received in Patients with Stage I Lung Cancer in Ontario: A Population-Based Analysis.

Current oncology (Toronto, Ont.)·2025
Same author

"Help with rowing the boat": Implementing and evaluating the Strengthening a Palliative Approach in Long-Term Care program in four Canadian provinces.

Palliative care and social practice·2025
Same author

The universal interprofessional education Q tool (U-IPEQ) for student learning- a pilot trial in the human anatomical dissection space.

BMC medical education·2025
Same journal

Methods for incorporating test result information within the high-dimensional propensity score framework: application in UK electronic health record data.

BMC medical research methodology·2026
Same journal

Sparse multi-way DMDC for longitudinal classification in high dimension low sample size data.

BMC medical research methodology·2026
Same journal

Tree-based exploratory identification of predictive biomarkers in non-randomized data.

BMC medical research methodology·2026
Same journal

Comparative evaluation of interrupted time series analytical methods for healthcare quality improvement research: a Monte Carlo simulation study.

BMC medical research methodology·2026
Same journal

Methodological advances in claims-based dementia algorithms: integrating medication and clinical data for medicare populations.

BMC medical research methodology·2026
Same journal

An interpretable XGboost algorithm for predicting 30-day mortality in acute pancreatitis using routine biomarkers.

BMC medical research methodology·2026
See all related articles

Fabricated medical data often shows higher correlation coefficients than real data. However, fabricated data may show smaller differences between group means, making scatter-plot analysis crucial for detecting research misconduct.

Area of Science:

  • Medical Research Ethics
  • Data Integrity in Science

Background:

  • Medical research misconduct is a growing concern.
  • Data fabrication is a severe form of misconduct.
  • Fabricated datasets may exhibit inflated correlation coefficients.

Purpose of the Study:

  • To investigate differences between real and fabricated datasets.
  • To assess the association between variables in fabricated versus real data.

Main Methods:

  • Compared fabricated datasets with real and simulated datasets.
  • Analyzed correlation structures of continuous variables.
  • Used independent t-tests to compare means between groups.

Main Results:

  • Fabricated datasets generally showed higher correlation coefficients than real datasets.

Related Experiment Videos

  • Even when real correlations were zero, fabricated data exhibited higher coefficients.
  • Fabricated data often displayed smaller or no differences between group means.

    • Conclusions:

    • High correlation coefficients (>0.70) can indicate data fabrication.
    • The rule of high correlation may not apply to differences between group means.
    • Scatter-plot analysis is a valuable tool for detecting fabricated data.