A big issue with the 2011 Census was that similar information was published across multiple tables but was described differently, due to the way the table was originally commissioned for a particular audience. As a user it meant that you had to grapple with sifting through the meaning to see if the tables were describing the same concept; you had to first find a table by its title – that sounded like it contained the information you wanted, then you had to open it and see if it contained the particular uk rcs data combination of variables that you were interested in. This repeated search and manual normalisation exercise was time-consuming and our users used to provide feedback saying they were never quite sure if they had identified all tables with the information they wanted.
Justin Hayes, Census support Manager tells us:
“So we thought that there must be a better way. Could the metadata be made consistent where all the things that are about the same concept are grouped together into topics, such as ‘age’ or ‘ethnicity’ and could the labels for variables within these topics be normalised? We looked around to see if there were standards that could be applied to the 2011 Census metadata which would enable this grouping and normalisation approach and we identified 2 possibilities: DDI and SDMX.”