News0 min ago
When Releasing Covid Infection Figures, Shouldn't They Also Release The Number Of Tests Carried Out?
This would give us a running daily actual proportion of tests against positive results.
Answers
//It’s all there if you want to blind yourself with statistics// It’s those who are blind to those statistics whom the government depends upon for its message. Some people who are a little inquisitive may find it extremely odd that on the day that the “truly astonishing” infection figures of 77k were announced, there were also the highest number of tests...
10:56 Fri 17th Dec 2021
I'm not convinced I've misinterpreted anything, PP; or at least no more egregiously than Gromit's post did. It depends, for example, on how the sample was taken.
I don't think there's any reasonable case for arguing, as Gromit does implicitly, that Covid is more prevalent now than a year ago, and his figures, presented as always out of context, are seriously misleading.
I don't think there's any reasonable case for arguing, as Gromit does implicitly, that Covid is more prevalent now than a year ago, and his figures, presented as always out of context, are seriously misleading.
Perhaps I should add that Gromit's particular mention of "now, with 50m vaccinated", with the clear tone that such programme has all gone to waste, is the particular part I'm picking up on. It has not gone to waste. Even setting aside the possibility that Omicron is a more mild variant, the benefits of the vaccination programme on reducing severe cases and deaths are manifest.
gromit: "They do release the figure of tests taken. Just nobody bothers to report it because it is meaningless. " - gawd help us, what school did you play truant from? What's meaningless is the number of positive tests without knowing how many tests were carried out. The OP is correct they need to publish the proportion of tests that are positive too.
// It's more likely to be sadly the inevitable consequence of getting older, as the chances of a mini-stroke increase as we age.//
James ! - you can check, estimate by looking at large samples.
als by doing special tests - which Khandy is NOT obliged to tell us if they were done or not
( and even if they were done, it has diagnostic significance. If no the doctors on the ground at the time did NOT think it was vaccine related)
( the strokes caused by covid and vacca re venous and sagittal, and the normals arent)
The fact that it has occurred includes it in ( the possibility) and does NOT exclude it out ( wd ve happened anyway)
You do very large prospective studies ( NHS very well placed for this as Big Pharma in the Land of the Free incredibly does NOT share data) - there is one for 300 000 does of vacca
and it turns out ( stroke after vacca) to be around 257 per billion. and You know how many strokes to expect in four weeks in a billion people with that age range - and the difference is due to vacca
and they - - - overestimated the vacca stroke rate ----
it is papers like this
https:/
even I read them with a 'jesus do I have to?' feeling
James ! - you can check, estimate by looking at large samples.
als by doing special tests - which Khandy is NOT obliged to tell us if they were done or not
( and even if they were done, it has diagnostic significance. If no the doctors on the ground at the time did NOT think it was vaccine related)
( the strokes caused by covid and vacca re venous and sagittal, and the normals arent)
The fact that it has occurred includes it in ( the possibility) and does NOT exclude it out ( wd ve happened anyway)
You do very large prospective studies ( NHS very well placed for this as Big Pharma in the Land of the Free incredibly does NOT share data) - there is one for 300 000 does of vacca
and it turns out ( stroke after vacca) to be around 257 per billion. and You know how many strokes to expect in four weeks in a billion people with that age range - and the difference is due to vacca
and they - - - overestimated the vacca stroke rate ----
it is papers like this
https:/
even I read them with a 'jesus do I have to?' feeling
//It’s all there if you want to blind yourself with statistics//
It’s those who are blind to those statistics whom the government depends upon for its message. Some people who are a little inquisitive may find it extremely odd that on the day that the “truly astonishing” infection figures of 77k were announced, there were also the highest number of tests reported since April and the fourth highest number since mass testing began.
Since the beginning of November (about the time the new variant was said to have arrived) the %age of positive tests among those reported has remained fairly consistent at between 4% and 5% (with occasional “outliers”). The average number of positive tests per 1,000 in that period is 44. On the day the 77K cases were announced it was 47. The previous Saturday it was 50. The number of positive tests is largely a result of the number of tests undertaken.
If 1.63m tests had been carried out last Saturday there would have almost certainly been a “truly astonishing” figure of around 80,000 new infections reported. If, when the figure of 77k new cases was announced, it was accompanied by a statement which said “This is almost twice as many as were reported on Saturday 4th December (41,457)” that would sound astonishing. If however if, instead, it was accompanied by a statement which said “This is almost twice as many as were reported on Saturday 4th December (41,457) but is hardly surprising because almost twice as many tests were undertaken (1,635,922 against 871,383) it would make much more sense and be far more useful.
//It's true that the testing data gets less emphasis than case data, but it's still easily available, requires barely any digging, and is probably more of a distraction than anything else.//
It’s not a distraction at all – it is fundamental to the story being presented. To major on “cases” when they have been established by testing is pointless unless you know something about the volume of testing. If double the number of tests were undertaken tomorrow (almost certainly producing double the number of “cases”) the (misleading) headlines would be “cases double in a day.” The volume of testing has a huge impact on the figures produced. We’re not talking about a percentage or two either way. 60% more tests were carried out last Wednesday than the previous Saturday (producing only 50% more cases so far from being blood-curdling the figures were encouraging). It is blindingly obvious that if you test such a greater number you will find a correspondingly greater number of positives and to not put that total into perspective is simply misleading. To turn it about, if we tested only 10% of the usual total (seeing a decline of 90% in positives) would it be open and honest to say “look how well we’ve done – cases down by 90% today”?
Once again, I await the figures showing the number of tests undertaken which produced yesterday’s 87k cases. But I don’t expect the government to provide any of the comparisons I just have because to do so would mean they have to devise another way to scare the population witless. Mass testing indicates that for the last six weeks around 95% of those tested do not have the virus. The only time cases have risen or fallen significantly is when the number of tests does likewise. It may well change, who knows? But to announce blood-curdling figures with no context is disingenuous.
It’s those who are blind to those statistics whom the government depends upon for its message. Some people who are a little inquisitive may find it extremely odd that on the day that the “truly astonishing” infection figures of 77k were announced, there were also the highest number of tests reported since April and the fourth highest number since mass testing began.
Since the beginning of November (about the time the new variant was said to have arrived) the %age of positive tests among those reported has remained fairly consistent at between 4% and 5% (with occasional “outliers”). The average number of positive tests per 1,000 in that period is 44. On the day the 77K cases were announced it was 47. The previous Saturday it was 50. The number of positive tests is largely a result of the number of tests undertaken.
If 1.63m tests had been carried out last Saturday there would have almost certainly been a “truly astonishing” figure of around 80,000 new infections reported. If, when the figure of 77k new cases was announced, it was accompanied by a statement which said “This is almost twice as many as were reported on Saturday 4th December (41,457)” that would sound astonishing. If however if, instead, it was accompanied by a statement which said “This is almost twice as many as were reported on Saturday 4th December (41,457) but is hardly surprising because almost twice as many tests were undertaken (1,635,922 against 871,383) it would make much more sense and be far more useful.
//It's true that the testing data gets less emphasis than case data, but it's still easily available, requires barely any digging, and is probably more of a distraction than anything else.//
It’s not a distraction at all – it is fundamental to the story being presented. To major on “cases” when they have been established by testing is pointless unless you know something about the volume of testing. If double the number of tests were undertaken tomorrow (almost certainly producing double the number of “cases”) the (misleading) headlines would be “cases double in a day.” The volume of testing has a huge impact on the figures produced. We’re not talking about a percentage or two either way. 60% more tests were carried out last Wednesday than the previous Saturday (producing only 50% more cases so far from being blood-curdling the figures were encouraging). It is blindingly obvious that if you test such a greater number you will find a correspondingly greater number of positives and to not put that total into perspective is simply misleading. To turn it about, if we tested only 10% of the usual total (seeing a decline of 90% in positives) would it be open and honest to say “look how well we’ve done – cases down by 90% today”?
Once again, I await the figures showing the number of tests undertaken which produced yesterday’s 87k cases. But I don’t expect the government to provide any of the comparisons I just have because to do so would mean they have to devise another way to scare the population witless. Mass testing indicates that for the last six weeks around 95% of those tested do not have the virus. The only time cases have risen or fallen significantly is when the number of tests does likewise. It may well change, who knows? But to announce blood-curdling figures with no context is disingenuous.
It would be lovely if I could read PP's post and know precisely what they were saying, since I'm never sure if he's agreeing with me or not. Still, we can but dream...
To reply to specific points PP makes: // Jim's first para misinterprets sampling - the bigger the sample the closer it is to the population figure ( of whatever it is you are measuring) //
Obviously. But the raw figures are still going to be undercounts, ie if you say "we looked at 10 fish in this pond of 100 fish, and found two blue fish, therefore there are two blue fish in the pond", this is obviously nonsense.
// His second para can be countered with ratios - in which case total number tested is relevant. This is called scaling by most of us (rates by NJ but he is an outlier, pun intended) //
Likewise I thought I had exactly explained this. The testing capacity has increased, meaning that the ratio is probably the more relevant measure; but on the other hand, since there are multiple types of test going on it's again extremely hard to extract any meaning out of the case/test ratio without further processing. So I didn't do it, beyond noting that an overall increase testing capacity is likely to be at least partially responsible for an increase in the recorded case figure between the two dates mentioned.
I don't claim to be the world's leading expert on Covid data and its interpretation, but I don't see that there's anything obviously wrong with what I posted, and it's certainly a better take than Gromit's usual out-of-context posts.
To reply to specific points PP makes: // Jim's first para misinterprets sampling - the bigger the sample the closer it is to the population figure ( of whatever it is you are measuring) //
Obviously. But the raw figures are still going to be undercounts, ie if you say "we looked at 10 fish in this pond of 100 fish, and found two blue fish, therefore there are two blue fish in the pond", this is obviously nonsense.
// His second para can be countered with ratios - in which case total number tested is relevant. This is called scaling by most of us (rates by NJ but he is an outlier, pun intended) //
Likewise I thought I had exactly explained this. The testing capacity has increased, meaning that the ratio is probably the more relevant measure; but on the other hand, since there are multiple types of test going on it's again extremely hard to extract any meaning out of the case/test ratio without further processing. So I didn't do it, beyond noting that an overall increase testing capacity is likely to be at least partially responsible for an increase in the recorded case figure between the two dates mentioned.
I don't claim to be the world's leading expert on Covid data and its interpretation, but I don't see that there's anything obviously wrong with what I posted, and it's certainly a better take than Gromit's usual out-of-context posts.
jim360 : "Secondly, if testing were performed in the same volumes, then this wouldn't matter, and you'd be able presumably to say that the correction factor to get the true figure would probably be the same. "
jim, what do you mean by "in the same volumes" ? - everyone? the same number of people? the same people?... or ?
NJ has no reason not to "cherry-pick" in his excellent post @17:00 Thu 16th Dec 2021 - https:/
as this statistical one-sided sham the Public have been presented since the start of this pandemic in verging on criminal.
jim, what do you mean by "in the same volumes" ? - everyone? the same number of people? the same people?... or ?
NJ has no reason not to "cherry-pick" in his excellent post @17:00 Thu 16th Dec 2021 - https:/
as this statistical one-sided sham the Public have been presented since the start of this pandemic in verging on criminal.
"In the same volumes" is roughly meant to mean "the same number of tests", but I hesitate to say it's exactly the same, because testing can be performed for many reasons. The most useful to track the true ebb and flow of cases is the surveillance testing, but this tends to be a couple of weeks behind so you don't get an up-to-date figure.
As LFTs can be taken every day by the same person, depending upon the scenario, how do you now if seventy million individuals have been tested in a week or just ten million?
Individuals can test negative daily and add to the negative figures but what about those who are positive, do they get tested every day to confirm they are positive still?
Individuals can test negative daily and add to the negative figures but what about those who are positive, do they get tested every day to confirm they are positive still?
//Obviously. But the raw figures are still going to be undercounts, ie if you say "we looked at 10 fish in this pond of 100 fish, and found two blue fish, therefore there are two blue fish in the pond", this is obviously nonsense.//
Quite so, Jim. But that's not what's being said. if you said "We looked at some fish in a pond of 100 fish two days ago found two blue fish and we looked at some fish in the same pond of 100 fish yesterday and found four blue fish, therefore the number of blue fish in the pond has doubled" that would be errant nonsense (and that is what's being foisted on the public). Unless you know how many fish were looked at the numbers are meaningless.
Variations in the number of positives mirror, very closely, variations in the number of tests reported and without the latter the number of cases is completely meaningless. You don't need to test all the population to establish the prevalence of the virus. More than 1m tests are carried out most days - more than enough to establish whether it is becoming more or less widespread.
There's obviously far more mining you can do on the data. However, the government is depending on simplicity. It is, however, being over simplistic. On Wednesday (and yesterday) it reported the "highest numbers ever of Covid cases." It did not say that on Wednesday (yesterday's testing figures are not yet available) virtually the highest number of tests (by a long chalk) were also carried out. Assuming the daily reports are on the same basis, it doesn't matter whether they include everybody, or some who have tested twice, etc. That bare (misleading) headline was enough to send people rushing to cancel their hospitality bookings for the Christmas period and send the hospitality industry into another crisis. It was enough for schools to close early and postpone the resumption in the new term.
I have a good idea: stop mass testing and the number of cases will reduced 99% overnight. A good idea? About as good as telling people that cases have doubled without telling them testing has doubled.
Quite so, Jim. But that's not what's being said. if you said "We looked at some fish in a pond of 100 fish two days ago found two blue fish and we looked at some fish in the same pond of 100 fish yesterday and found four blue fish, therefore the number of blue fish in the pond has doubled" that would be errant nonsense (and that is what's being foisted on the public). Unless you know how many fish were looked at the numbers are meaningless.
Variations in the number of positives mirror, very closely, variations in the number of tests reported and without the latter the number of cases is completely meaningless. You don't need to test all the population to establish the prevalence of the virus. More than 1m tests are carried out most days - more than enough to establish whether it is becoming more or less widespread.
There's obviously far more mining you can do on the data. However, the government is depending on simplicity. It is, however, being over simplistic. On Wednesday (and yesterday) it reported the "highest numbers ever of Covid cases." It did not say that on Wednesday (yesterday's testing figures are not yet available) virtually the highest number of tests (by a long chalk) were also carried out. Assuming the daily reports are on the same basis, it doesn't matter whether they include everybody, or some who have tested twice, etc. That bare (misleading) headline was enough to send people rushing to cancel their hospitality bookings for the Christmas period and send the hospitality industry into another crisis. It was enough for schools to close early and postpone the resumption in the new term.
I have a good idea: stop mass testing and the number of cases will reduced 99% overnight. A good idea? About as good as telling people that cases have doubled without telling them testing has doubled.
If you test 30% of chickens and turkeys in the supermarkets for traces of salmonella you would no doubt get 29% infection rate, test 60% the positive test would go to 59%, test 80% then you would get 79%. If you forecast the above you destroy the poultry industry for no good reason. Unless your daft enough to eat the poultry raw you have nothing to fear
The percentages of positive PCR tests in England are shown in this link,
https:/
https:/
LFD(LFT) data methodology for England is given here,
https:/
https:/
Latest Posts
ChatterBank10 mins ago
Can First Class Stamps Be Traced
ChatterBank12 mins ago
Listener 4842 For Me And You By...
Crosswords14 mins ago
Crosswords37 mins ago
Related Questions
Sorry, we can't find any related questions. Try using the search bar at the top of the page to search for some keywords, or choose a topic and submit your own question.