DRAG*N
A C R O
automatic checking of research outputs
Webinar Wednesday 17th November 2021, 14.30-15.30 GMT (15.30 – 16.30 CET)
Checking research outputs for residual disclosure risk requires skilled analysts to carry out manual checks on statistical findings. In 2019 a team at University of the West of England, Bristol (UWE) was commissioned in by Eurostat (Methodology; Innovation in Official Statistics Unit) to develop a proof-of-concept.
The resulting code, ACRO (Automatic Checking of Research Outputs), is a software tool designed and developed to reduce the manual workload of statistical disclosure control (SDC) checking of research outputs. With simplicity, readability and low training needs in mind, the authors demonstrate the feasibility of partially automating output checking, thus envisaging a potential reduction of the resources allocated. ACRO currently allows the safe tabulation of outputs and the computation of several popular estimators. At present it is only coded for Stata®, version 16 or earlier.
This webinar will give a demonstration of ACRO and then discuss the project: its genesis, its aspirations, its limitations, the initial hopes and the lessons learned. We intend that most of webinar time will be given over to Q&A so that we can explore these issues and identify useful development routes.
The team are particularly keen for more users to provide ideas and feedback. Users wanting to try the software beforehand can see the demonstration video at [link] and download the code, manuals and test data from http://www.fivesafes.org/SDC/ACRO/
This event is suitable for anyone responsible for, or interested in, statistical disclosure control of research outputs based on confidential data. The webinar will be hosted by Elizabeth Green, Felix Ritchie and Jim Smith of UWE, with support and comments from Eurostat.
To register for the event, please click here.
The resulting code, ACRO (Automatic Checking of Research Outputs), is a software tool designed and developed to reduce the manual workload of statistical disclosure control (SDC) checking of research outputs. With simplicity, readability and low training needs in mind, the authors demonstrate the feasibility of partially automating output checking, thus envisaging a potential reduction of the resources allocated. ACRO currently allows the safe tabulation of outputs and the computation of several popular estimators. At present it is only coded for Stata®, version 16 or earlier.
This webinar will give a demonstration of ACRO and then discuss the project: its genesis, its aspirations, its limitations, the initial hopes and the lessons learned. We intend that most of webinar time will be given over to Q&A so that we can explore these issues and identify useful development routes.
The team are particularly keen for more users to provide ideas and feedback. Users wanting to try the software beforehand can see the demonstration video at [link] and download the code, manuals and test data from http://www.fivesafes.org/SDC/ACRO/
This event is suitable for anyone responsible for, or interested in, statistical disclosure control of research outputs based on confidential data. The webinar will be hosted by Elizabeth Green, Felix Ritchie and Jim Smith of UWE, with support and comments from Eurostat.
To register for the event, please click here.