Data are fundamental to research: theories and their ultimate acceptance or rejection are based on them. But there is concern in the academy that many researchers lack the skills needed to manage their data, which could include anything from photographs or interview recordings to genome sequences.
To address the problem, the UK Data Archive (UKDA), based at the University of Essex, is running a series of workshops to help.
"The focus is on sharing and ensuring that data are usable in the long term," said Veerle Van den Eynden, research data management support services senior officer.
It is hoped that by improving researchers' data-management skills, information will be shared more widely and thus increase the value of British research.
Dr Van den Eynden outlined to Times Higher Education the key points for researchers.
First, preparation is vital. She said that by introducing a data-management plan at the start of a project, researchers can make their lives easier in the long run. Many funders already insist on this.
She also encouraged researchers never to assume that data are unsuitable for sharing. Potentially sticky issues such as ethics, confidentiality and copyright should be considered case by case, she said.
Where consent for data sharing is necessary, she recommended that it be requested openly during collection.
She also pointed out that it is often helpful to assure participants that information will be shared only with other researchers, not the general public.
Dr Van den Eynden was keen to dispel the perception that by sharing data, academics lose control over their research outputs. Conditional access arrangements allow them to retain right of approval and can even initiate new collaborations, she said.
As an additional precaution, she suggested that sensitive data be stored in repositories under embargo, as the Government does.
The workshops focus on the use of databases. Dr Van den Eynden said that when researchers create them, they must consider how others can benefit from their work.
She said: "They should ask themselves: 'How can I describe the information so that others in my team and beyond can understand it?'"
This could mean documenting variable names and explaining codes or abbreviations. It is also helpful to include details of the methodology used in data collection and the contact details of original researchers, she added.
Another tip is that if data are created in collaboration with others, keeping track of the different versions produced is important - either via specialist software or by maintaining a record of what changes are made, when and by whom.
Dr Van den Eynden pointed out that there are numerous ways to store data. Most universities have their own repositories, as do many research councils, but other services are available, including advice from organisations such as the Edinburgh-based Digital Curation Centre.
While shared data are stored almost exclusively in digital form, often the raw material is on paper, which throws up additional considerations.
Institutions subscribing to international guidelines must ensure that records "remain legible, readily identifiable and retrievable", according to the International Organisation for Standardisation, with documents kept for a minimum of seven years.
Although the aim of the workshops is to improve data sharing, Dr Van den Eynden said that researchers should be careful not to deposit or share data sources for which they do not have the copyright.
"You're allowed to use data in the public domain in non-profit research as long as you acknowledge this," she said, but you cannot pass the data on.
Bere Mahoney, a psychologist at the University of Worcester, said the UKDA course had shown her how valuable data sharing could be.
"Simply searching data sources is in itself an exciting starting point for the development of new and existing projects," she said.
For more information, visit: http://www.data-archive.ac.uk/sharing/dmstraining.asp
FIVE STEPS TO SUCCESS
Plan: address issues of consent and check funders' data policies at the outset
Record: decide the best format in advance and keep track of your data-generation process
Document: keep a catalogue of metadata and details of variables and methodology
Store: archive hard copies securely, identify the most appropriate electronic repository and make sure everything is backed up
Share: check the copyright of your materials, identify which aspects of your information are sensitive or confidential, then approve access procedures.