NOTES FROM THE FOCUS GROUP MEETING HELD ON MONDAY 8TH OCTOBER
ABOUT THE FOCUS GROUP
The Focus group was carried out as one of the normal meetings of the Nottingham Café Scientifique et Culturel on the evening of Monday 8th October 2012. This society meets for the purposes of ‘public engagement’ with the latest ideas arising in science and culture. The audience is mainly comprised of academics, professionals, and students. The audience therefore has an interest in understanding research and associated matters. As the ‘general public’ they are also interested in how public money is spent on research, and what happens to the outputs that are gained from this research.
Prior to the Focus Group starting, the purpose of the Focus Group was explained to the participants. They were then asked to sign a consent form and given a sheet of suggested questions which they could refer to throughout the Focus Group. They were asked to provide either their experiences or opinions related to this area.
The following topics arose during the course of the Focus Group Meeting.
THE RESEARCH OR DATA OF THE FOCUS GROUP ITSELF
The focus group reported a variety of academic, practical (e.g. professional purposes such as obtaining community data to submit funding bids), and personal research projects (e.g. database of output of personal research interests in an academic field). The focus group participants seemed to have a data sharing mindset and overall felt that data should be shared.
LOCATION OF THE DATA / LOCATING DATA
People wondered where data should be submitted so that it did not get lost – this is important as it is a public record produced often at the public’s expense.
What is the best method of finding data?
Journals – People still publish via journals, people are used to this model, and it means that people then know where to look for research output.
Use of Google Scholar – Google Scholar can help with locating studies. But Google itself provides a search list which shows the items that are most frequently consulted, rather than necessarily showing those which are of better quality.
Institutional Repositories – However, these are not consistent from one organisation to another; they have different methods and the software can be configured differently. IRs may lead to searching Google instead.
WHEN DO YOU SHARE RESEARCH DATA – AT WHAT POINT IN THE PROCESS OF RESEARCH?
At what point in the research process should the data be shared?
Should there be a choice about the timing of release?
Raw Data – Should the data be in its ‘raw’ state or should it be contextualised by the researcher first? The data in some of its early states may not be comprehensible or usable by others. In these states it could be liable to misuse. It may be better to release the data once it is determined that there are no errors in it which could lead to unreliable studies by other researchers.
Interpretation before release – If people are still processing the data, they may feel the need to interpret it before sharing it. They may thus wait until the PhD, or other report is finished, before going public with the actual data. People would not necessarily want to share their data prior to producing their publications in order to maximise the number of publications.
The nature of the data – It may depend on what people want to use the data for, and the nature of the data itself as to whether shared data is useful. Is this more of an issue for qualitative data which is based on the interpretation of the researcher, rather than quantitative data?
Relevance of the data – Should the data be released while it is still of interest? Old data may lose its relevance or appeal.
Peer Review – The data should be available for the review process to enable peer-review to check the data. This could however be a time consuming process. Not all reviewers may feel they have the time to check the data as well as the article to which it relates.
BENEFITS OF DATA SHARING
New outcomes – Other people may be able to produce fresh interpretations of the data to advance the subject. Different researchers may find patterns that other people have missed.
Preservation – data which is copied and updated by others is more likely to be preserved; it is also more likely to be checked and is thus more reliable.
Ensuring reliability – e.g. making pharmaceutical data open ensures that it is not ‘rubbish’ (see arguments of Ben Goldacre)
Producing a sharing culture – everyone sharing their data means that people cannot ‘bury’ flawed research.
Collaboration/Comprehensivity – sharing a personal database of research means that other people would be able to contribute; one person cannot collect all the data necessary for the project. This would then lead to a comprehensive database.
Pooling data – sharing data would enable data to be pooled from different sources.
ISSUES WITH DATA SHARING
Confidentiality – Issues of confidentiality were raised related to data sharing which would make it difficult to be shared.
Infrastructure – Lack of infrastructure in the researcher’s organisation may deter data sharing.
Preservation – Data formats: some are not straightforward; digital data may have been stored in formats that are no longer used (floppy discs for example); more reliable formats are needed; readers for obsolete data types may be required. What would assist with data preservation? (e.g. more reliable formats such as tablets of stone, the web).
Time – It is time consuming to prepare data for sharing.
Value judgments – Who is qualified to make a judgment on what data should be preserved, as not everything can be preserved?; Who should have the job of filtering other people’s minds? Will this lead to value judgments being made about some forms of data?
Knowledge is power – it is also access to future funding. People may be concerned about sharing data if it means that it is used by others in a way which prevents them from obtaining future funding to continue with the line of research.
Misuse – future analyses may be incorrect, or cherry-picking of the data may take place to aid a particular argument – and data which does not support the argument can be ignored.
Processed data – people may claim that the data has been fiddled with (processed in some unreliable way).
Lack of Knowledge of how to share data – Someone reported that they did not know how to share data but would like to be able to do this.
Information Overload – A data sharing culture may mean that eventually there is too much information out there to manage successfully.
New research – Research could become a process of analysing old datasets rather than producing new data. Science would then become a process of interpretation.
Different languages – This could be a barrier to collaboration and sharing.
Ownership disputes – There could be disputes between authors as to who owns the data.
Verification studies – Funders do not want to fund them, they are seen as low status and not worthwhile. Journals do not want to publish straightforward replication studies; they value newness but this does not mean that the study is necessarily worthwhile. Again there are value judgements being made here, but not by the researchers themselves. The researchers are at the joint mercy of funders and publishers.
New models of data sharing – The way data is shared changes frequently (e.g. CD v iTunes model); people have to keep up to date with the environment of sharing.
Financial models – publishers need to make money, may impede the process of data sharing? OA needs to find a way of being sustainable.
Data Citation – This ensures that all data re-use is cited so that the original researcher(s) get(s) the credit for the data they have produced.
How to incentivise? – Given that University promotion is based on new research and high impact journals, how can researchers be incentivised to share their data if they perceive that this may weaken their professional progress?
Peer review – Peer review of data could lead to public attributions of merit.
LEVELS OF ACCESS
Free information – One attendee wanted to make their personal research available – but wanted the access to be free.
Researcher pays? – This seems like vanity publishing to one of the attendees.
QUANTITATIVE V QUALITATIVE DATA
Someone mentioned that they could not find statistical data to back up their research but anecdotal, qualitative research supported their assumption. If they had waited for the supporting figures it would have taken too long to set their community project in motion. This is why community groups are now commissioning research.