When setting up an institutional RDM Service, you quickly realise that there are many unknowns: How many RDM queries will you receive each month? What type of questions will they ask? How much time will I need to allocate? This blog post has been written for anyone currently scoping their institutional support service and looking for real world examples on which they can base their estimates. It outlines work performed by the LSHTM RDM Support Service to help researchers with their data management questions during 2013.
What is an RDM query?
First, we need to decide where to establish the boundaries of a RDM query. For simple queries, this is easy; a researcher emails a question and you send a reply. For complex queries, it’s a bit trickier. If a researcher submits several questions on Monday, then another set on Wednesday, are these part of the same query? If the second set are unrelated to the first, I’d say yes, but what if some are related and others aren’t? If a series of correspondence cover various topics, but you realise that they’re being entered into a DM Plan, do you re-classify them into a single query?
When preparing monthly reports, I’ve taken a simple approach: a reported query covers all interactions with a person during a stated month. The discussion may cover a range of topics, but will be part of the same query. If a query is submitted on the last day of the month and discussion continues a week later, it’s included in the report for the month it was completed. This has worked for the RDM queries I’ve received so far, but may change if correspondences are more frequent.
How many queries have been received?
I should begin with an admission: we spent relatively little effort promoting the advisory component of the RDM Support Service during 2013. Sure, we mention it at dissemination events, but we haven’t actively gone out and promoted the fact that we provide advice. I continue to be surprised that people get in touch and ask questions, and even more amazed that I’m able to answer them. Hopefully, this will be addressed in the next few months, through creation of new posters/leaflets and the forthcoming RDM website.
In total, we received and responded to 88 unsolicited queries in 2013. Forty-five of these were submitted by researchers in our Epidemiology and Population Health (EPH) faculty. The Public Health and Policy (PHP) and Infectious and Tropical Diseases (ITD) both submitted seventh queries. A further six queries came from external academics, mostly researchers who were collaborating on LSHTM projects. There were also a small number of correspondences with people in Academic and Administrative Services and the Division of Education.
It’s difficult to identify trends at this stage. The end of term and Christmas period were reasonably quiet this year (I was on holiday so some queries were processed later).
The high points coincide with training events, drop-in sessions and workshops being organised by the RDM Service. Having dealt with several panicked researchers, I’m aware that proposal submission deadlines are also a contributing factor for when some people get in touch, but this isn’t evident on the bar chart. Most of the advice I’ve provided has been for active projects.
How many researchers have contacted the RDM Service?
A second method of analysing the information is to count the number of researchers who have submitted queries during 2013. For this, I’m only looking at the primary contact for each query. Although they may bring a colleague to a meeting/CC others into emails, the researcher count would increase significantly, making it meaningless.
In total, 68 people contacted the RDM Service during the year. By comparing the number of queries with the number of researchers, we can see that only 20 queries were made by people who had previously contacted the RDM Service. The majority of researchers contacted the RDM Service once only and did not get in touch at a later date. We don’t have sufficient information to draw conclusions from this, unfortunately. They may be satisfied with the response and simply not require further help, or utterly traumatised by the experience. It may be useful to perform some form of follow-up on previous contacts to see if they’ve encountered other problems.
How much time is spent on each query?
I’ve made only limited progress in determining the average time that should be spent on an RDM query. I’m not keeping a record of the exact time spent on each query, but find that most take between 30-60 minutes. Easy-to-answer queries or those that are similar to another query can be processed in 15 minutes. At the top of the scale, reviewing Data Management Plans and technical specifications can take up to 4 hours of time. Much of this is spent meeting with the researcher and answering questions.
Once LSHTM’s Research Data Management website has officially launched, I expect the queries I receive will change. For general questions, I’ll be able to direct them to the relevant section on the website. However, I will likely need to spend more time providing advice on how recommendations may be applied by projects in specific situations.
What type of questions are raised in an RDM query?
The questions raised in an RDM query vary significantly, depending upon where the researcher is in the research process. Topics covered include the following:
- Writing a Data Management Plan
- Data Management costs
- Locating datasets for use in teaching & research
- School policy on Data Sharing
- Data Transfer Agreements and consent forms
- Data Manager recruitment
- Access to 3rd party data
- Advice on specific software (e.g. OpenClinica)
- Data capture using mobile devices and web-based tools
- Database design
- Data normalisation and cleaning
- Data Encryption
- Data visualisation
- Locating suitable data repositories to manage and share data
- Processes for curating and preserving data within School
Some concluding thoughts
It’s difficult to quantify the impact of the RDM Support Service at this stage. I think we’ve done a good job in improving awareness of data management processes in the research process. Although most researchers consider data management to be important, many felt overwhelmed by the amount of information available and were unaware of many developments in this area. During the next year I hope to build upon what we have achieved so far, by running various dissemination events and working directly with research projects to improve RDM practices.