The Problem of Biased Data and Potential Solutions for Health and Environmental Assessments

ABSTRACT

The utility and credibility of environmental assessments depend on the use of unbiased data. However, it is increasingly clear that, despite peer review, much of the scientific literature is biased. Sources of bias include publication practices, research design and implementation, funding influences, investigator expectations, statistical methods, confounding, suppression, and fraud. Assessors can take precautions against biased data by performing their own reviews of the sources of data, checking for retractions and corrections, requiring full disclosure of methods, acquiring original data and reanalyzing it, avoiding secondary sources, avoiding unreplicated studies or studies that are not concordant with related studies, and checking for funding or investigator biases. Journals, government agencies, and other institutions can take many more types of actions to reduce bias in scientific data. Some of these are already being implemented but others will require a greater willingness to enforce scientific ethical standards.

Key Words:

• research bias
• data
• risk assessment
• research fraud
• peer review
• data quality

View correction statement:

Corrigendum

ACKNOWLEDGMENTS

We thank Rick Ziegler, Harlal Choudhury, Peter Chapman, and two anonymous reviewers for their helpful comments and thank participants in the National Center for Environmental Assessment, Cincinnati, Science Meetings for lively discussions of these issues. The views expressed in the article are those of the authors and do not necessarily represent the views or policies of the U.S. Environmental Protection Agency. The authors declare that this work received no support other than the salaries provided by their employer and they have no conflicts of interest. However, they do have a bias in favor of good science.