This post is about the first part of a three-part briefing on the responses to our survey carried out in May 2015. The briefing is available on our survey page. A basic report giving analysis of these responses was sent to the survey participants at the end of June. We promised not to disclose the identity of respondents or their institutions. This meant we could not include results in that first report if they were potentially disclosive. The questions on staffing levels were in that category, and given that these may be of interest for planning purposes we have used these as the topic of the first briefing. The next will focus on resourcing (questions 10 & 11); and the third on where institutions retain data (question 9) and when they will have the capability to ensure long-term access (question 16).
According to our responses as of May 2015, the average allocation to the central RDM service was 1.7 FTE per institution. That is expected to rise to 2.0 by May 2016. There is likely to be some noise and uncertainty in the institution’s responses. We did not get a high enough response rate to be very confident about generalising. On the other hand from our direct experience of working with institutions to support their development we often find this is being taken forward as part of one or two people’s existing responsibilities. And we might say ‘thank heaven for small increases’ to any increase in staffing in the present climate. But is it enough, and what implications do the responses have for the RDM sector in the UK?
Whether the level of staffing is ‘enough’ obviously depends on what kinds and level of capability an institution’s central RDM function ought to be providing. Let’s assume that a reasonable compliance-driven service would offer researchers some regular training and advocacy sessions, provide advice on data management plans, give pointers to best practice advice, manage access to active research data storage on an institutional network, and help them get their data deposited into an institutional repository, or outwards to external data centres. That roughly corresponds to the scope described in our guide How to develop RDM services.
There were responses to the staff allocation questions from 52 institutions. To get a better picture of these results they were analysed by the institutions’ research income and intensity. A breakdown by whether institutions were EPSRC funded or not did not make sense for this question as all but 8 responses were from those receiving funding. There are a few noticeable patterns in the responses, which suggest:
- At least two thirds of institutions currently have less than 1 FTE allocated to RDM
- There is a striking gap between ‘research income rich’ and ‘research income poor’ institutions; the richest third expect to have almost 3 FTE on average by May 2016, while the poorest third will still have less than 1 FTE.
- There is high variation around the mean, indicating that institutions are taking into account a wider range of factors than research income and intensity in their RDM staffing allocation.
- Library, Research & Enterprise Office, and IT service staff will each continue to play a role in the ‘business as usual’ RDM service.
- A drop-off in temporary IT staff is expected, perhaps in line with a move from project-based to ‘business as usual’ service.
I’ll comment further below on the survey results and suggest institutions need to be sure their RDM services are capable of changing quickly. We still need better information on researcher demand, and from where they expect it to be met. Community-wide shared services will change the support available, alongside rapidly developing commercial services that institutions need to integrate with. However our results suggest most institutions have allocated a bare minimum of staff time to deal with this.
According to our responses as of May 2015, the average allocation to central RDM service development and delivery was 1.7 FTE per institution. That is expected to rise to 2.0 by May 2016. There is likely to be some noise and uncertainty in the institution’s responses. We did not get a high enough response rate to be very confident about generalising. On the other hand from our direct experience of working with institutions to support their development we often find this is being taken forward as part of one or two people’s existing responsibilities. And we might say ‘thank heaven for small increases’ to any increase in staffing in the present climate. But is it enough, and what implications do the responses have for the RDM sector in the UK?
Whether the level of staffing is ‘enough’ obviously depends on what kinds and level of capability an institution’s central RDM function ought to be providing. Let’s assume that a reasonable compliance-driven service would offer researchers some regular training and advocacy sessions, provide advice on data management plans, give pointers to best practice advice, manage access to active research data storage on an institutional network, and help them get their data deposited into an institutional repository, or outwards to external data centres. That roughly corresponds to the scope described in our guide How to develop RDM services.
Through DCC’s institutional engagement programme we are aware of many RDM services that have been driven forward on the efforts of one or two people. However across the sector as whole I we need a better understanding of what level of capability is being offered, and what volumes of enquiry or ‘case load’ those services are dealing with. That is something we plan to address in the coming months.
My own impression is that institutional RDM services do not vary much in terms of what they support, but do vary a great deal in the level of capability offered, the extent to which their services are developed in-house, and the level of responses they are getting from researchers. Again it is only an impression, but it seems to me that most are so far only really scratching the surface of potential demand.
How can we estimate that demand? Numbers of academic staff at least support back of the envelope calculations. According to HESA’s contextual data for the REF there were about 92,000 staff submitted to the 2014 REF assessment. Between 152 higher education institutions, that represents an average of just over 605 per institution. If each of those staff only requested support from the central RDM service once per year that would be about 3 enquiries per FTE to deal with each day. That’s assuming one FTE translates to an optimistic 200 productive working days, each of which involves nothing else but responding to support requests. That is only one aspect of the service. It doesn’t factor in postgrad students who may be eligible for support, or any attempt to limit the service scope to research supported by particular funders.
Some institutions use much more sophisticated models than that to plan their service. But the community’s collective ability to plan may be limited by lack of good data. Institutions tend not to share their business-planning models with their competitors. DCC can probably help here by gathering examples of the factors and assumptions used (if you think you could help here please get in touch!). We are planning case studies, aiming to get institutions beyond the ‘we need to be seen to be doing something but can’t push up our overheads’ model.
A surer picture of how researchers expect to deal with RDM would also help. Will all the awareness-raising that is going on persuade them to wholeheartedly buy in to their central institutional RDM service? Or is it more likely they will get together with their peers elsewhere to develop something they think better meets their sub-disciplinary needs, perhaps using cloud-based tools they can easily adapt?
We might learn from how institutional computing services support for research group online activity has developed over time. I suspect most research groups and individuals expect only a minimal profile of their activity on the corporate web presence. But if they prefer to ‘do their own thing’ in terms of web-based research output, gearing this towards peers and users of their own research niche, that is often with a great deal of back-office web support from central or faculty-level services.
The central RDM service may need to remodel itself quickly, whether to respond to dec-centralised activity or the spectrum of cloud-based RDM tools services developing beyond the institution. Research groups should be able to turn to the institution to provide more sophisticated back office functions, at least archiving, persistent identification and discovery layers, integrating these into researcher-owned presentation and analysis layers. There will be a lot of institution-specific metadata to gather and make sense of, however other services evolve. It will be very interesting to see the RDM shared service that Jisc is piloting will take shape.
Services are evolving quickly, and there could be big pressures on central RDM services to rapidly adapt the scale and complexity of their offerings. Shared services that can be delivered on a cloud or hybrid basis will no doubt help address that. But will these fit the needs of the many institutions with only a single FTE? RDM service development can involve a lot of ‘change management’. The institution needs to be able to apply that to the RDM service itself as well as to its academics’ data practices.