Jayne,I'm afraid I must disagree - the rise in the number of tests is important in understanding the real number of cases.As an example, let's say you have a bucket containing 100 balls, some red, some green but you don't know the numbers. Pick out 10 balls, 3 of them are red. You may infer that the most likely proportion of red balls is 30%. Now pick out 20 balls, 6 of them are red. I hope you will see that the proportion in the second sample (6/20) is the same as the first (3/10) i.e. both 30%. I think that according to your reasoning you would want to say that the real number of red balls inferred from the second test is twice than that inferred from the first. I say the inferred real number is the same.(The number of balls is the total population, the red balls are infected cases, each pick of a ball is a test)There are complications. The value inferred is not a number, it is a distribution - scientists often quote a 95% confidence interval - so in our example we might say we are 95% confident that the real value lies between (say) 25 and 35. (I'm afraid my stats is too rusty to work this out properly)Even worse, there is the problem of false positives and negatives. This becomes serious where the reliability of the test is poor (e.g. say out of 100 positive tests only 80 are really positive) and the real prevalence is low (say 1%). In this case, the false positive can swamp the actual positives.Now, all of these points may have been covered in the reporting of the Leicester case, but if so I haven't seen it.With these sources of noise in the data I am not convinced that the authorities are justified in imposing further lockdown on this city with its known negative economic, psychological and health effects.
Andy Riley ● 2316d