More often than not, science experiments in student labs fail. We tell ourselves that in “real” labs, things are done differently and experiments will work. However, in a scientific community after the South Korean cloning fraud, there is an increasing concern that research findings we trust to be true are false. John Ioannidis of the Tufts University School of Medicine explains this in an article boldly titled “Why most published research findings are false.”

Statistics are used most often to validate research findings, but Ioannidis uses statistics to throw cold water on scientific research in general. The post predictive value, PPV, is the probability of a finding being true after a study has been done. Microarrays and other high-throughput techniques, many of which have revolutionized biological research, have an extremely low PPV.

When coupled with the issue of “effect sizes,” the importance of the finding to everyday human life, research findings begin to look even shadier. Ioannidis says the findings that least affect human life will be plagued with “ubiquitous false positive claims” while scientific fields with large effects, like smoking and cancer, are more likely to publish true findings.

Ioannidis deduces several interesting corollaries about the probability of a research finding being true. For example, the greater the financial interests and prejudices in a scientific field, the less we can trust their claims.

“The hotter the scientific field (with more scientific teams involved), the less likely the research findings are true,” says Ioannidis. When “timing is of the essence in beating competition,” experts actually suppress new findings that refute established findings through the peer review process.

Every day, novel nomenclature is coined to describe new scientific discoveries. At the same time, colorful jargon dealing with the less glorious side of research is generated. The Proteus phenomenon, for example, refers to the playground battle of alternating claims and refutations. Scientific teams often refrain from publishing negative results until another team finds a positive result for the same question and publishes it in a prestigious journal.

Ioannidis recommends the use of statistical analysis of a given study in the wider context of similar studies to improve the current situation.