I thought the question WR asked was whether ONS data should be presented in that way? NS arrived at her statement that England had 20% higher infection by taking the 2 infection rates (something like 4.5% in Scotland and 5.5% in England) and (correctly) stated that 5.5 is 20% higher than 4.5. However, she didn't actually state the overall percentages (4.5/ 5.5) or the ratio of the population (approx 1 in 20 people in both Scotland and England) as the ONS had. WR pointed out that that was a bit twisty and not telling the whole story, which wasn't the best use of statistical data to make a political point.
The uksa pointed out that it is better to provide the full info, as the public, understandably have difficulty distinguishing between percent and percentage points, and that in practice at that point the infection rates were pretty much the same in Scotland and England. Had the percentages been, for example, 20% and 24% there would also be a 20% difference, but that would be a much more pronounced difference... something like 1 in 5 set against nearly 1 in 4 people...
Uksa also responded to Jackie Baillie's complaint that JS was using pre-Christmas infection levels to show how brilliant the Scottish Post Christmas restrictions were. While acknowledging that JS was at the time using the most up to date data, they did say...
When using statistics in public communication it is important to be clear about what the data are measuring, and the choice of time period is all the more important in a rapidly changing situation.
Which is pretty relevant to both complaints, really...