My answer is even simpler: Pretending that lockdown was a sharp cutoff on March 24th is ahistorical bullshit and should be laughed at rather than analysed.
19th March - people told to stop all nonessential contact
20th March - schools close, pubs/clubs/theatres told to close early and remain closed
21/22 March - weekend, so workplaces were shut
23rd March - official full lockdown announced
24th March - date used as the ‘lockdown date’ in bullshit analyses.
26th March - lockdown legally comes into force
So schools were closed by Government order four days ahead of the ‘lockdown date’, entertainment venues ~3.5 days ahead of it, and workplaces were closed for 2 out of 3 days preceding the ‘lockdown date’. That fully and entirely explains any minor discrepancy.
Just like the actual legal changes that came into effect on the 26th, the lockdown announcement on the 23rd was a recognition of events that had already happened.
I agree with Peter that there were measures in place before official lockdown on 24 March, and this should have been mentioned by Wood et al.
An anti-lockdown argument might be that "If people were limiting contact before official lockdown, then doesn't that prove that compulsory lockdown was unnecessary?", but I don't think this is right. For one thing, as Peter points out, some of the pre-24 March measures were compulsory (closed schools), and as Oliver points out in his note, even a slight increase in R would have lead to a lot more deaths. And there is the point (mentioned at point 4 in my RSS response note) that people react in anticipation of compulsory measures. E.g., the PM announced on 16 March that “now is the time for everyone to stop non-essential contact and travel", which many people will have taken as an order. And the schools were closed from 20 March, but the announcement of this was made on 18 March. (I remember taking my daughter to school on 19 March and it was like a ghost town.)
I largely have to agree with this. My wife’s then employer moved to home working the Wednesday before lockdown, and they weren’t the first to do that in our area. Also, companies that didn’t move to home working did make changes to practices that would have reduced transmission.
Hence, and I am not saying that Oliver himself is making any particular statement on lockdown, even, if infections peaked before the “official” lockdown date, this should not be used as evidence that such interventions do not work or are unnecessary
Sure, and I don't think anyone is seriously taking the 26th as the date for example. But I think it's a worthwhile exercise to analyse the curve properly, because if it did turn out that infections peaked on the 18th say (and I don't think they did) then that would have been useful information
To an extent, sure - you’d still have interminable arguments about whether ‘voluntary’ behaviour changes were durable, or just anticipation of what we could all see coming.
But the analysis is now done, and people are still trying to argue that the difference between peaking on the 21st vs the 24th is meaningful. It’s not.
Of course we can *debate* whether it would still have peaked if we had allowed workplaces and shops to remain open and stopped at closing schools, entertainment and hospitality (ie status as of end of 20th March).
But there is zero point trying to use the real world data to argue the point, because the only relevant information we have is one highly unusual working Monday on the 23rd. You would have to resolve transmission dynamics to within less than 8 hours to answer the question. Sat/Sun 21/22 are not informative.
I more interested if there was any requirement for a compulsory lockdown that covered what people did in their own homes and when outdoors in public spaced. Eg could businesses and organisations have just been restricted?
I saw the bit about how a narrow curve in infection space gives a wider curve in death space and was thinking about the uncertainty principle and Fourier transforms and I see that’s roughly where the analysis went.
Deconvoluting simple things is hard so something like a pandemic will be very tricky. Should be taken seriously but not to within days.
Ultimately I think the important function of lockdowns - as you say - is to get some decent time at lower r and so maximise the benefit of the lockdown.
I'm undoubtedly getting a self-selecting sample via screen shots, but I don't have the impression that genuinely held gender-critical views are treated respectfully over there for example. And there are various other hot button issues that I wouldn't go within a mile of over there but would happily do so on Twitter
Quite possibly! But I find Twitter just too degraded now and the signal/noise is just not worth it. Read the odd bit of football news and that’s about it.
My answer is even simpler: Pretending that lockdown was a sharp cutoff on March 24th is ahistorical bullshit and should be laughed at rather than analysed.
19th March - people told to stop all nonessential contact
20th March - schools close, pubs/clubs/theatres told to close early and remain closed
21/22 March - weekend, so workplaces were shut
23rd March - official full lockdown announced
24th March - date used as the ‘lockdown date’ in bullshit analyses.
26th March - lockdown legally comes into force
So schools were closed by Government order four days ahead of the ‘lockdown date’, entertainment venues ~3.5 days ahead of it, and workplaces were closed for 2 out of 3 days preceding the ‘lockdown date’. That fully and entirely explains any minor discrepancy.
Just like the actual legal changes that came into effect on the 26th, the lockdown announcement on the 23rd was a recognition of events that had already happened.
I agree with Peter that there were measures in place before official lockdown on 24 March, and this should have been mentioned by Wood et al.
An anti-lockdown argument might be that "If people were limiting contact before official lockdown, then doesn't that prove that compulsory lockdown was unnecessary?", but I don't think this is right. For one thing, as Peter points out, some of the pre-24 March measures were compulsory (closed schools), and as Oliver points out in his note, even a slight increase in R would have lead to a lot more deaths. And there is the point (mentioned at point 4 in my RSS response note) that people react in anticipation of compulsory measures. E.g., the PM announced on 16 March that “now is the time for everyone to stop non-essential contact and travel", which many people will have taken as an order. And the schools were closed from 20 March, but the announcement of this was made on 18 March. (I remember taking my daughter to school on 19 March and it was like a ghost town.)
Yes, I agree - I think the other side of it is that since growth did undoubtedly slow prior to full lockdown, people have to be careful making claims about "delaying lockdown by a week cost X lives" https://www.theguardian.com/world/2020/jun/10/uk-coronavirus-lockdown-20000-lives-boris-johnson-neil-ferguson - I don't think you can just assume the 3-4 day doubling continued (and half the lilypond was covered at the last minute)
On March 18, The Times splashed on "Britain in Lockdown".
I largely have to agree with this. My wife’s then employer moved to home working the Wednesday before lockdown, and they weren’t the first to do that in our area. Also, companies that didn’t move to home working did make changes to practices that would have reduced transmission.
Hence, and I am not saying that Oliver himself is making any particular statement on lockdown, even, if infections peaked before the “official” lockdown date, this should not be used as evidence that such interventions do not work or are unnecessary
Sure, and I don't think anyone is seriously taking the 26th as the date for example. But I think it's a worthwhile exercise to analyse the curve properly, because if it did turn out that infections peaked on the 18th say (and I don't think they did) then that would have been useful information
To an extent, sure - you’d still have interminable arguments about whether ‘voluntary’ behaviour changes were durable, or just anticipation of what we could all see coming.
But the analysis is now done, and people are still trying to argue that the difference between peaking on the 21st vs the 24th is meaningful. It’s not.
Of course we can *debate* whether it would still have peaked if we had allowed workplaces and shops to remain open and stopped at closing schools, entertainment and hospitality (ie status as of end of 20th March).
But there is zero point trying to use the real world data to argue the point, because the only relevant information we have is one highly unusual working Monday on the 23rd. You would have to resolve transmission dynamics to within less than 8 hours to answer the question. Sat/Sun 21/22 are not informative.
The link to Alex Selby's response seems to be to David Spiegelhalter's contribution.
Thanks, I've fixed that now (too many links!)
I more interested if there was any requirement for a compulsory lockdown that covered what people did in their own homes and when outdoors in public spaced. Eg could businesses and organisations have just been restricted?
I saw the bit about how a narrow curve in infection space gives a wider curve in death space and was thinking about the uncertainty principle and Fourier transforms and I see that’s roughly where the analysis went.
Deconvoluting simple things is hard so something like a pandemic will be very tricky. Should be taken seriously but not to within days.
Ultimately I think the important function of lockdowns - as you say - is to get some decent time at lower r and so maximise the benefit of the lockdown.
Personally I find Bluesky free from nonsense
I think how pleasant the experience is might depend whether your views fit within a certain range or not
Possibly. But genuinely held opinions are treated a lot more respectfully than in the other place.
I'm undoubtedly getting a self-selecting sample via screen shots, but I don't have the impression that genuinely held gender-critical views are treated respectfully over there for example. And there are various other hot button issues that I wouldn't go within a mile of over there but would happily do so on Twitter
Quite possibly! But I find Twitter just too degraded now and the signal/noise is just not worth it. Read the odd bit of football news and that’s about it.