Because it is full of bias and somewhat unreliable despite pretence to the contrary (and a heap of money spent). I discussed the sorry state of this effort way back - if I can find my posts again I'll send a link!
I recall that what had me swearing early on with the national survey was the sample frame - they recruited folk whose contact details they already had from previous surveys and then decided to recruit folk who thought they'd had COVID! They then contracted an organisation that flipped the whole thing to quasi-random half way through which is why the rates change. They then decided to change the sample method again. At no point was there a long series random sample or stratified sample. Oh, and they decided to target 'hot areas' when they felt like it, for which purpose they had mobile lab. Pig's ear is not the word!
Oddly enough, I work for one of the research agencies sub-contracted by ONS, and I am a capable and committed (over ten years experience) field research interviewer. I took one week to try working on this study in autumn 2020 before dropping it like a hot potato: I won't go into all the tedious details, but it all looked very flawed and rushed and the management communications were very poor.
I didn't stick it long enough to be in a position to criticise the methodology, I simply 'had a bad feeling about it'. I'd be interested to read your own critique if you could locate the link.
I found the "5 month lag" in data late last year, but called it a "seasonal" effect.
How you checked with data in nations with delayed vaccination of their working age population to see if their lag is still 5 months, or if it's 3.5 or something that frames the health burden with the extreme seasonal months (Summer for the first round of jabs)? When I looked at data, summer jumped out at me more than did an exact clock.
Whahey! Cracking! I'm not deluded, then! :-) I've not taken a look at any international data for quite a while now since I simply don't have time. Hopefully Joel, Jason and others will already have cracked that.
Absolutely Rob. When I was the suit writing the big cheques for national surveys I only considered bids with significant prelim work and an interim report - that's where the contract breaker was put just in case the whole thing fell apart. Omnishambles is a perfect description!
Because it is full of bias and somewhat unreliable despite pretence to the contrary (and a heap of money spent). I discussed the sorry state of this effort way back - if I can find my posts again I'll send a link!
Yes, it was a while ago after I sat with their methodology document. I'll have a rummage. We should start a G7 club!
Rounds will be cheap then!
I recall that what had me swearing early on with the national survey was the sample frame - they recruited folk whose contact details they already had from previous surveys and then decided to recruit folk who thought they'd had COVID! They then contracted an organisation that flipped the whole thing to quasi-random half way through which is why the rates change. They then decided to change the sample method again. At no point was there a long series random sample or stratified sample. Oh, and they decided to target 'hot areas' when they felt like it, for which purpose they had mobile lab. Pig's ear is not the word!
Oddly enough, I work for one of the research agencies sub-contracted by ONS, and I am a capable and committed (over ten years experience) field research interviewer. I took one week to try working on this study in autumn 2020 before dropping it like a hot potato: I won't go into all the tedious details, but it all looked very flawed and rushed and the management communications were very poor.
I didn't stick it long enough to be in a position to criticise the methodology, I simply 'had a bad feeling about it'. I'd be interested to read your own critique if you could locate the link.
In your first slide the first CDR peak looks to be 27, yet in later slides it’s about 13. Why? Am I missing something?
Different time frame!
Ah yes. Sorry
Always good to query my work - I've made some right bloopers!
I found the "5 month lag" in data late last year, but called it a "seasonal" effect.
How you checked with data in nations with delayed vaccination of their working age population to see if their lag is still 5 months, or if it's 3.5 or something that frames the health burden with the extreme seasonal months (Summer for the first round of jabs)? When I looked at data, summer jumped out at me more than did an exact clock.
Whahey! Cracking! I'm not deluded, then! :-) I've not taken a look at any international data for quite a while now since I simply don't have time. Hopefully Joel, Jason and others will already have cracked that.
My eyeballs fell out. Searching to be ready for next slides.
Absolutely Rob. When I was the suit writing the big cheques for national surveys I only considered bids with significant prelim work and an interim report - that's where the contract breaker was put just in case the whole thing fell apart. Omnishambles is a perfect description!