Needed Improvements

Peer review today is a flawed system distorted by subjective opinions, personal biases and incentives. And the process often doesn’t work. Several research areas are experiencing a replication crisis and often major flaws in papers are only pointed out after publication. Alarmingly, two to thirty-four percent of published papers may be frauds, depending on the field and the source. These can be generated by paper mills, automated gibberish paper creators, or researchers plagiarising others and faking data. The British Medical Journal, for example, ran experiments that deliberately put errors into papers and sent them out to the standard reviewers, who missed 25% to 30% of them, including major flaws. Furthermore, almost USD 200 billion (85%) of annual global spending on research is wasted on badly designed or redundant studies.

In addition, unimportant papers are still published. The numbers vary, but numerous research suggests the number of papers that are never cited once is quite large. For example, one source reports that 82% of papers go uncited in the humanities, 27% in the natural sciences, 32% in the social sciences, and 12% in medicine. Editors and/or reviewers also struggle to identify duplicate submissions because of the gated nature of academic journals. This results in flawed and rejected papers being published. In sum, the academic peer review system is in crisis, and the costs are extreme.

“When prospective employers look at our CVs, it is much easier for them to judge the quantity of our output over the quality of the research.” - physics lecturer from a prominent Beijing university

Academics collectively spend 15,000 people years reviewing ~4.7 million articles every year for publishing in 30,000 scientific journals. That is a tremendous amount of talent to waste in a flawed system. In addition, the gated silos of academic journals are expensive, costing cash-strapped universities millions in subscriptions.