(Horia Varlan via Flickr)

Scientists around the world are constantly creating or collecting data in the course of their research. But a study by a group of academics led by Timothy H. Vines at the University of British Columbia suggests that those data often become  inaccessible in the years after research is published.

The researchers took a sample of over 500 studies published between 1991 to 2011, then made an effort to contact the authors and request a copy of the original data. The researchers only managed to track down and receive data for 19 percent of the papers studied. In most cases, the researchers were unable to contact paper authors. Other authors replied that their data still existed, but they were unwilling to share. And when the author did respond, models based on their answers show that the odds of the requested data set still existing fell by 17 percent per year.

This is a big deal because public funds support a significant chunk of academic research — and those data sets could be used to inform research down the line. But Vines and his fellow researchers say their findings reinforce the notion that individual researchers aren't the most reliable group to be charged with preserving those data sets in the long term. Instead, they suggest policies mandating that data be shared via public archives maintained by research funders or publishers.

And they're not the only ones who think a shift of that nature might benefit the larger scientific community. For instance, a White House directive this year aimed at increasing access to the results of federally funded research included a call for broader access to digital data sets in addition to peer-reviewed published results.