The power of open access (OA) publishing to enhance the impacts of research is highlighted by a new study carried out by scientists in Canada and the UK and published in the journal PLoS ONE. The researchers hope their findings will encourage more universities, research institutions and research funders to adopt OA self-archiving mandates, in which all research published under their auspices is made freely available to other researchers.

Articles that are made freely available in OA repositories tend to be cited more often than similar articles that can only be viewed by paying subscribers. The reasons for this have been the subject of much debate.

To get to the bottom of the matter, Professor Harnard and colleagues in Southampton and at the Université du Québec à Montréal in Canada studied papers from the first four institutions in the world to make OA mandatory for their researchers - these were Southampton's School of Electronics and Computer Science, CERN (the European Organization for Nuclear Research), Queensland University of Technology in Australia and Minho University in Portugal.

The team compared the citation impact of papers from these institutions with comparable papers produced by institutions with no OA mandate; in some cases the authors had opted to make their articles OA anyway; in most cases, they had not.

The investigation confirmed the impact of OA on citations; Open Access articles were cited significantly more than non-OA articles, even when other variables are taken into account. Furthermore, the open access advantage is just as great for articles that have to be OA because the authors' institutions require it as it is for articles that are open access at the authors' choice.

This means that OA articles are not cited more often because their authors only make them OA if they think they are more citable. Rather, through open access, higher quality articles are made more accessible and therefore more citable. Had they not been made OA, these articles would only have been available to people with a subscription to the journal, making them less accessible and so less citable.

The article was published during Open Access Week, an international initiative designed to promote the open access cause and provide OA's advocates with an opportunity to share knowledge and ideas.

Some 2.5 million articles are published in 25,000 peer-reviewed journals and conference proceedings every year. Currently, just 15%-20% of these articles are self-archived in OA repositories. However, as the authors of the paper point out, no institution can afford to subscribe to all the journals in the world that its researchers might need.

However, growing numbers of institutions and research funding organisations are starting to put in place requirements regarding open access. For example, the European Research Council (ERC) requires all peer-reviewed papers arising from ERC-funded projects to be placed in appropriate repositories and made open access for a maximum of six months after publication.

The European Commission is also running an OA pilot project, under which projects funded under 7 areas of the Seventh Framework Programme (FP7) must be deposited in an online depository and made open access 6 or 12 months after publication. The pilot is set to run until the end of FP7 and if successful, it could serve as a model for future framework programmes.