Physical scientists use and access information in very different ways depending on the precise field they work in, according to a report released today by the UK's Research Information Network, the Royal Astronomical Society and the Institute of Physics, which publishes physicsworld.com. Google Scholar, for example, is used by 73% of Earth scientists and by 70% of nanoscientists to discover new research findings, but by just 13% of particle physicists and 7% of astrophysicists. Meanwhile, whereas all chemists and Earth scientists surveyed say they read online journals, only 38% of particle physicists do so, largely preferring preprint servers such as arXiv.

Entitled Collaborative Yet Independent, the new report is based on interviews with 51 researchers and focus-group sessions with 35 participants in seven different fields. It reveals that although physical scientists have led the way in using computers to analyse data, they are still fairly conservative when it comes to adopting new communications technologies, with formal publication in traditional journals "remaining the gold standard" for disseminating findings. Indeed, the report says that talking to peers and experts seems likely to remain one of the most important ways for such researchers to learn about new results.

Few physical scientists use blogs, Twitter, Open Notebook Science, social networks, public wikis or other "public-facing" technologies to share research information, the report finds, although some particle physicists and astrophysicists use internal, private wikis. Most physical scientists view these services as "distractions" from their communications with key colleagues – the only exception being researchers involved in "citizen-science" projects such as Galaxy Zoo, which rely on close collaboration with members of the public. Indeed, three-quarters of particle physicists still use e-mail lists to find new information.

Another issue the report highlights is the unwillingness of physical scientists to reference scientific databases, despite such scientific data increasing in volume and becoming ever easier to access. "There is little agreement on how to cite databases, or otherwise assign credit to the scientists and technicians responsible for the creation and maintenance of databases," the report says. But finding ways to assign credit is important, it adds, because otherwise "those responsible for creating data have fewer career incentives to engage in such efforts".

Monica Bulger from the Oxford Internet Institute, who co-authored the report, says that one problem with getting scientists to change how they access information is that they tend to have picked up the "tools" of their field in the lab as graduate students and later only learn new techniques in response to the needs of a particular project. "I wouldn't classify senior figures as frowning on new techniques and tools, but they may be less likely to experiment for the sake of trying something new," she says. Bulger was also surprised by the "fragility" of scientific research, which depends on "shrinking sources of funding and the use of often outdated computing systems due to lack of financial support".