Imperial College Data Audit Framework Implementation: Final Report

TitleImperial College Data Audit Framework Implementation: Final Report
Publication TypeReport
Year of Publication2009
AuthorsJerrome, N., & Breeze J.
Date Published2009/02/26/
InstitutionImperial College London
Keywordssg_data_survey, sg_ps

Gap Area Study Type:

To pilot the Digital Asset Framework Methodology; evaluate the scale and scope of research data; and make recommendations accordingly
Used a modified form of the Digital Asset Framework in multiple departments: used the audit framework in a first phase of investigation, then conducted an online survey and follow up interviews.
The pilot audit was useful in identifying current data management practice in the department and in agreeing recommendations. However the DAF method was found to be mostly unsuitable for identifying, classifying and managing data assets. The main reasons were: a) not aligned with current methods for categorising data; b) the large amount of time and researcher effort required to populate the register; and c) the importance of ‘non data’ assets such as computer software. As a result, an alternative approach was agreed for the remaining assessments. An online survey was developed and used as the primary data gathering method, supported by 1:1 interviews with a small number of selected researchers. The overall conclusion of the subsequent assessments was that the survey approach is also an imperfect technique. If it is completed by co-ordinators within research groups, the complexity and heterogeneous range of data is difficult to capture on a survey form. On the other hand if it is distributed to all researchers, a low response rate (typical for surveys) does not provide confidence in the extent of coverage. Online surveys are only likely to produce the desired results if mandated or motivated to complete (e.g. as part of a research proposal, or required as part of College research practice). A further conclusion was that a ‘one size fits all’ approach to research data management is not appropriate due to wide ranging nature of data (size, type, longevity, degree of collaboration), even within the limited sample evaluated by the project.