We were delighted with the response to our recent survey on research data management training in the sciences – almost two hundred researchers took the time to complete the questionnaire.
The respondents came from a wide variety of fields, with medical sciences and life sciences being particularly well represented, and from all career stages through from graduate students (who made up about half the survey sample) to senior researchers.
The survey focused on eleven key data management tasks. For each of these, we asked researchers how confident they felt about it, how much training (formal and informal) they had received, and how beneficial they felt additional training would be.
The research we’ve done previously has indicated that most researchers have received little formal training in research data management. However, some anecdotal evidence suggested that science researchers were perhaps more likely to receive informal on-the-job training within their labs or research groups.
The survey results confirmed this. For all eleven tasks, researchers reported receiving more informal training (which might take the form of advice from colleagues or supervisors, or being directed to written guidance material) than formal training (such as taught courses, or exercises on which feedback was given).
However, overall training levels were generally fairly low. There was only one task out of the eleven (managing bibliographic data) for which at least a quarter of respondents had received more than a little formal training. There were five tasks for which at least a quarter had received more than a little informal training, but none for which more than half had.
Perhaps unsurprisingly, the tasks in which most training had been received were those relating to day-to-day management of information: managing bibliographic data; organizing and structuring data within files; storing data securely and backing up; and organizing, structuring, and naming files and folders. Those for which least training was reported were those which concern what happens to data after the end of a project: preparing datasets for long-term preservation; determining whether datasets ought to be preserved; preparing datasets for sharing with other researchers; and dealing with copyright, licensing, or IP issues.
It also came as no real surprise to find that the two tasks researchers felt least confident about – dealing with copyright, licensing, and IP issues, and preparing datasets for long-term preservation – were also the areas in which they felt training would be most beneficial. Correspondingly, the tasks for which additional training was regarded as least necessary were those about which researchers felt most confident: organizing, structuring, and naming files and folders, and managing bibliographic data. In between these extremes, however, there were some more unexpected results. For example, version control was a task that received a relatively low confidence rating (joint fourth lowest out of eleven), but was also one of the areas in which additional training was viewed as less useful (ranked ninth out of eleven). Conversely, storing data securely and backing up ranked relatively highly in terms of both researchers’ confidence (fourth highest out of eleven) and training received (third highest, for both formal and informal training), but was still in the top half of the list of tasks (joint fifth highest) when it came to respondents’ feelings about the usefulness of additional training – perhaps reflecting researchers’ awareness of how essential it is to keep their data safe.
Reassuringly, respondents showed a substantial degree of enthusiasm for additional training in this area as a whole. One researcher commented that data management had never been mentioned as a subject before, despite having managed data on two large projects. This respondent concluded ‘I think I’ve managed OK but I was probably reinventing the wheel.’ This seems to reflect many researchers’ experiences of frequently being left to work out how to do data management unaided – with little or no guidance about best practice or ways of making the process easier. Another respondent simply commented ‘I’m pleased to see that this area has been recognised as one that requires formal training.’
For those who are interested in knowing more, a spreadsheet summarizing the full survey results (with a few illustrative charts and graphs) can be downloaded from the DaMaRO website. We’d also be happy to supply a copy of the raw underlying data to anyone who’d like to carry out further analysis of their own – please email us on firstname.lastname@example.org
The survey results have given us a useful snapshot of the training science researchers are currently receiving and would like to receive – and thus have provided some helpful pointers about where best to focus our efforts over the coming months.