At a recent speech in Virginia, Attorney General Jeff Sessions proclaimed that undocumented immigrants in Arizona were more likely to go to jail than U.S. citizens. His source was a study by the Crime Prevention Research Center.
But there was a problem. That study had a fatal flaw that undermined its conclusion. In fact, the majority of literature on the subject actually shows the reverse.
This isn't the only example of an analysis gone awry. In fact, faulty research is common in all the sciences. And that's particularly worrisome in the medical sciences.
Faulty science has dire ramifications. Every year, our nation funnels billions of dollars into research that is then used to make crucial policy decisions with even larger implications in lives and dollars. It's imperative that science give us facts, not false impressions.
That's where the emerging information technology known as blockchain could come to the rescue. By authenticating and certifying published research data using the blockchain, the scientific community could reduce errors and regain the public's confidence.
Roughly 2.5 million studies relaying innovative discoveries and solutions are published every year. These analyses are backed by huge investments. The National Institutes of Health alone invests over $32 billion annually in studies conducted at universities and research institutes worldwide.
But those studies can rarely be authenticated. Over two-thirds of scientists who have tried to replicate the results of other experiments have been unable to do so. So, if a research paper incorrectly classifies data, makes a quantitative analytical error, or reaches biased conclusions, there's no obvious method to identify and correct that mistake.
Researchers themselves admit that the results of many studies are unreliable.
Take the words of Richard Horton, editor-in-chief of The Lancet, one of the world's most renowned medical journals. Horton acknowledged that up to half of all published scientific work could be wrong. He attributes those inaccuracies to "invalid exploratory analyses" and "flagrant conflicts of interests," among other factors.
This has terrifying real-world implications. Consider cancer research. In 2012, researchers evaluated 53 renowned cancer studies that had been published in top-tier journals. In the end, only six of them could actually be replicated.
That's scary stuff. Four in ten Americans will develop cancer in their lifetimes. Researchers need to be able to tell good science from bad as they search for the next generation of cures—and they need to be sure their own research is on target.
Scientists already have the ability to share raw data prior to the publication of their studies—but have few venues and little incentive to do so. Blockchain would make sharing, authenticating, and certifying information much faster and easier—and render it nearly impossible to alter data.
Blockchain offers a secure chain of data in what's called a "peer-to-peer" network. Any member of the network can access this shared data and use the blockchain to exchange information. When a "block"—a unique set of data—has been added, everyone with access to that chain can see the change.
A blockchain network would open a revolutionary window to the verification of research. Rather than release a data set upon publication of a scientific article, researchers could upload their data as soon as they collect it. A timestamp on the data would mark it as their own, preventing others from stealing it—and settling any argument over who made the discovery first.
Voila, a process that used to be completely closed to outside eyes before publication would now be open. Put simply, blockchain would serve as a massive vessel through which we could track records, verify data, and replicate results.
That would be a step up from our current system. Right now, getting a study published involves a "peer-review" process. So experts briefly read the study, make suggestions, and recommend whether it should be published. But with blockchain, every part of the scientific process would be transparent. This type of distributed authentication model would help restore trust in the scientific process.
Researchers may not embrace blockchain on their own immediately. But if funding groups, especially the NIH, support this approach and insist on the use of blockchain as a prerequisite for research funding, scientists would come to accept it. Top scientific journals, too, could require all published studies to incorporate blockchain.
And down the road, as blockchain becomes more widely accepted as the first-rate credibility standard, researchers could bypass the publisher middle-man and post results online directly—enabling the scientific community to share information more quickly.
Science is full of faulty practices. That has dire consequences for individuals and science as an enterprise. Blockchain can enable scientists to question faulty research right away, and make sure facts are facts.