Should I be more open?
In UK Higher Education we talk a lot about being open with our data, our policies and our impacts. We also talk about the need for transparency (and are rightly criticised when we’re not) and more recently about our civic roles, working with and to support our local communities. But we do really know what all this means and are we putting any of it into practice?
I’ve worked in a host of different roles at 3 different Russell Group Universities and an HE in FE College. Like most others in the sector, we had some pretty clear rules of what was expected of our academics and their research outputs.
Each had policies outlining the difference between green open access (open, delayed) and gold (paid for but immediately available), which extolled the general benefits of open access. Though rarely did they use something as visual as Jisc’s diagram below:
Thinking about some of my own jobs, particularly those related to research such as a European Commission (EC) funded project, which I managed as part of their Framework Programme 7, there were often clear requirements that our data and reports be made accessible. The EC openly encouraged ‘authors to publish articles via immediate open access’.
As the Jisc diagram alludes to this had benefits for the academics as well as the funders. By sharing their data it would likely mean more citations; more people being able to view, engage with and utilise their work; their results being more widely known and best case scenario — this would ultimately lead to other breakthroughs either in their discipline or others. From a funders perspective this equally meant not having to potentially pay for the same research more than once.
Making all data and publications open and accessible is admirable but it did sometimes cause us problems. Our data was often too big to host on University servers and some journals required us to pay to allow the article to be given immediate open access (and we didn’t have enough funding to do this for all publications), while others set embargoes so we couldn’t publish it on our own websites or in our own repository before they had published it themselves.
But despite it being pretty standard practice to at least try to make your research data and publications as open as possible, I can’t remember a single institution having a clear strategy or policy on which elements of internal data, policy or processes could (or even should) be made publicly available.
Rather than meet things like the OKFN’s definition that data should be available for anyone to ‘freely access, use, modify and share for any purpose’, the default was —and at least in my experience — still is, not to share.
I have to admit I’d never really properly considered the distinction in the expectations of different categories of staff before and am thankful to the PG Cert in HE (the reason behind writing this blog) for giving me the prompt!
What’s stopping us being more open?
In line with the frascati definition of research, Universities want to improve and extend knowledge of man, culture and society. To suggest we aren’t open at all with our internal data would be a falsehood, fake news if you like. We do make a large amount of data available. Granted, this is usually through mandatory annual returns to the Higher Education Statistics Agency (though there are plans afoot to make some of these returns a more frequent occurrence through the data futures project, but perhaps we shouldn’t talk about how well that’s progressing).
We publish data on our students, staff, estate, how many businesses we engaged with over a given period of time and a plethora of financial details. This data is all publicly available and is used by Universities and journalists alike to better understand the health of the sector and of individual HEIs (Higher Education Institutions).
Though we know the benefits of sharing, there are also some pretty robust arguments to suggest we shouldn’t, and indeed sometimes can’t be wholly open with our data or internal findings:
Consent and ethics:
- Student and staff data we hold is difficult to truly anonymise, meaning sharing this could result in our finding ourselves in breach of GDPR. Even when anonymised, we need explicit consent to share the data from the data subjects and we may not always know how the data could be used once it is open.
- If we were to remove all fields which have the potential to identify individuals, in some cases this would actually limit the usefulness of any analysis.
- The intention of those wanting to use data isn’t captured when making a source open. In the most part this would not be a problem, but we should ask ourselves what if the data was used inappropriately, or inaccurately to influence government officials and/or policy.
Data cleansing and formatting:
- Internal data often requires a lot of cleansing and formatting before it can be made useful. There is a cost to this which would need to be met by the institution. Given current uncertainties around University finance this resource requirement would need to be balanced amongst other priorities.
Competency of analysts:
- In our new ‘post truth’ world, there is an obvious risk to opening data up to false interpretation by those who do not have the requisite expertise or skills
And we’re definitely not as open to sharing when it comes to our methodologies and formulae for say planning student numbers; or the volume of research applications we submit each year to secure x number of projects or y level of income (both of which are more widely known); or what work each of our staff are engaged in to give an idea of productivity (though there are some projects between groups of Universities to break this down) — and I’m sure the list could go on.
The reason behind this, though, can be parked at the door of recent efforts to commercialise higher education. Rightly or wrongly, this has led to HEIs acting more corporately and seeking to retain whatever commercial advantage they can — you only need look at jobs.ac.uk and the number of data based roles advertised each week to see how much weight is being given to this type of work.
OK, but are these good enough reasons not to be more open?
The Office for Students (OfS), our new regulator would I imagine, say not. Just over a year ago, the OfS opened as our new regulator, marking a clear shift away from the previous role HEFCE had, as more of a critical friend. OfS have openly put students at the heart of their work and one of their key requirements of registering (which HEIs have to do in order to be called a University and for their students to be able to access student loans), is for Universities to meet conditions of transparency.
This is especially important for our conversations with students and has led to Universities across the country reviewing how they portray data such as how we spend tuition fees or fund new initiatives, as you can see from these sites:
This also shows that there are other ways that Universities can contribute to open knowledge without necessarily releasing all of the raw data. University policies and reports can (and are) being more readily shared on the public facing side of webpages, instead of the password protected intranets. This means we can all share good practice, saving time and resources in not reinventing the wheel.
We know that Universities are always looking to make savings and to run themselves more efficiently, particularly in these times of unprecedented uncertainty. Yet, collectively (and individually!) we spend a lot of our resources on evaluating, analysing and planning, often using the same sets of data and trying to second guess the methods or aims of our competitor institutions.
Much has been said about the commercialisation of higher education and one outcome of this is the fear of commercially sensitive data getting into the hands of our rivals. But this goes against everything a University stands for.
So, should I be more open?
Yes.
Sharing information across the sector (and beyond) both in terms of data and our processes and policies would benefit a whole host of stakeholders:
- Our students: because we would better understand their needs by being able to analyse larger sample sizes and potentially pick up trends that weren’t visible at just one institution)
- Our staff: because we don’t all need to replicate the same policies or documents, by sharing this information more freely we can reduce the duplication of effort and concentrate instead on adding value)
- Our wider communities of interested parties: because they would be able to interrogate our data or hold us accountable for how we allocate our resources)
