Based on data from Altmetric is supported by Macmillan Science and Education, which owns Nature Publishing Group.

Search engines have revolutionized how scientists find papers — especially articles that have been around for a while. A team of researchers at Google has documented a surge in the citation rate for older papers. Scientists had a range of responses online. The study found that 36% of citations in 2013 were to papers that were at least 10 years old — a 28% increase since 1990. Carlos Baquero, a computer scientist at the University of Minho in Braga, Portugal, tweeted:

The authors say that the digitization of journal archives and online search engines have made it easier than ever to find older papers.

“Now that finding and reading relevant older articles is about as easy as finding and reading recently published articles,” they write, “significant advances aren’t getting lost on the shelves.”

The authors also showed that citation rates for papers from the past have climbed sharply in many fields, including the humanities, social sciences, physics, life sciences and computer science. However, for reasons that aren’t clear, the trend is much smaller in engineering, chemistry and material sciences.

Computer scientist Lance Fortnow of the Georgia Institute of Technology in Atlanta, who shared the Google study with his followers on Twitter, said in an interview that putting scientists in touch with previous studies can help to stop them from “reinventing the wheel”.

But he also sees some downsides. “Many breakthroughs by young researchers happen because they don’t have old ideas getting in the way,” he says. “If old thinking is too easy to find, it might be too hard to ignore.” Fortnow adds that he misses the serendipity of accidentally finding interesting papers while rummaging through libraries. “When you use Google Scholar, you only find what you search for.”

Kiran Garimella, who studies information technology at Aalto University in Espoo, Finland, tweeted:

Garimella told Nature that Google Scholar is “screwing up” citations by ranking search results on the basis of the number of citations that a paper has already received. “This creates a loop where highly cited papers are cited more because they appear in Google Scholar,” he says. He believes that the focus on highly cited papers is especially counterproductive in computer science, where new and better algorithms are created at a fast pace. In the end, he says, cutting-edge research gets short shrift. Still, “Google Scholar is an amazing tool and I wouldn’t have been able to do a lot of things without it,” he admits. “Part of the problem is with the laziness of researchers who don’t always bother to look beyond the first page of results — myself included.”

Anurag Acharya, co-founder of Google Scholar and a co-author of the latest paper, counters that the search engine doesn’t create a bias towards older papers. He notes that Google Scholar allows users to restrict their search to papers from recent years. “This is by far the most-used search feature,” he says. He also points out that search results can be sorted by date, so researchers who want to focus on the latest papers can do so. “We have put in substantial effort to make it easy for researchers to find new research,” he says.