Ideally we would have waited until all the data had been released by the agencies before beginning processing to reduce the numbers of changes we had to make to our working model. This would have meant that we would have been better placed to look at all of the data to see how best to describe the concepts. However, we wanted to release the data as soon as possible to allow our users to access it.
Richard Wiseman,
“We identified an issue with how we had treated ‘Economic activity’. This codelist had doubled in size between our previous release (which primarily contained ONS data) and this new release. The larger america rcs data codelist now contained about 200 codes. This was due to erring on the side of caution when we described the data. We decided to investigate this to see if we could normalise this codelist further, as we knew from our download logs that it is the Economic activity codelist most often downloaded by our users.
During this investigation, we discovered an issue to do with whether students were counted as economically active or inactive. We followed these through by looking in detail at the data to determine whether the numbers were the same even if we’d opted to describe something slightly differently. We discovered that in a lot of instances, this was in fact the case, which meant we could reduce the numbers of codes used. We also decided to reorganise the code hierarchies to make it more explicit for our users to see whether the figures included students or not. This then had a knock-on effect in that we were able to match more data, which then meant we needed to redo the append step creating new versions of our composite UK tables.”