Talk

Advanced search

Why aren't these figures being reported?

(46 Posts)
IdrisElbow Sat 25-Apr-20 18:51:26

This graph has been produced by a man called David Paton, a professor of economics at Nottingham University. Link to his twitter here:

twitter.com/cricketwyvern/status/1254041723526922241?s=19

He is plotting the deaths by date of occurrence rather than by date of reporting. The government numbers quoted each day are by date of reporting not by date of occurrence. This is giving a consistently bleak picture of very little improvement. However, if you go by numbers 'reported' there is a lag time of several days and even weeks. For example, today's figures encapsulate deaths that happened two weeks ago.

If you look at the graph by the date deaths are recorded, you can see a clear downward trend. I am just interested to understand why the stats aren't being shown more widely in this way. Am I missing something? (I hope I've explained that clearly, it's taken me a while to get my head around it all 😬)

OP’s posts: |
dementedpixie Sat 25-Apr-20 21:33:29

I thought we already knew the peak day was 8th April. I think there have been a few graphs like the one you have shown

Drivingdownthe101 Sat 25-Apr-20 21:34:15

I’ve seen this graph (or similar ones) quite a few times

IdrisElbow Sat 25-Apr-20 22:30:03

Ah right, ok, entirely possible I've missed it in the mainstream media. I only listen to the daily update on the radio rather than watch it in tv so not sure if this is what they are showing on the slides? It just seems to me to give a different picture if the situation compared to the 'deaths reprted' figure that is released every day.

OP’s posts: |
LilacTree1 Sat 25-Apr-20 22:32:02

Just asked for alternative stats on main thread as I’m so annoyed by this. They just want to make it look bad. Like Patrick Vallance said himself, many with covid on their death certificate were not tested.

IdrisElbow Sat 25-Apr-20 22:43:14

Yes I agree, it makes for a far more dramatic story to report the 'reported deaths' figure. However, in the context of CV where we are looking for a pattern, the date of death is far more important. Especially when the 'reported deaths' data is so skewed by the fact that the lag time for reporting some deaths is very large, weeks in some cases. The very least the Gov and media should do is report both sets of the data.

OP’s posts: |
LilacTree1 Sat 25-Apr-20 22:49:48

Yes OP

But then they have no excuse for their lockdown.

Sammy867 Sat 25-Apr-20 22:52:32

The nhs release the statistics every day and back date the numbers. If you only follow the news you will always get skewed information, depending on how they want the statistics to sound to the public. To be honest they could publish the nhs updated graph every day but they don’t and I don’t really see why as their method of publishing the numbers is ridiculous and tells us nothing.

This image is todays. The orange are the cases announced today ( you can see some go back two weeks) and the blue are previous announced. You can see the trend that 8th was the peak and numbers are slowly dropping

Cuddling57 Sat 25-Apr-20 22:53:23

I agree with you. Why isn't this graph on the government daily briefing?
Are you reading the mumsnet thread 'Daily numbers, graphs ...'? Very interesting and the OP does a volcano graph with the real dates on.
I feel like I should say 'these aren't just statistics ... each one is very sad 😢'

IdrisElbow Sat 25-Apr-20 23:15:46

@Cuddling57 I am not on that thread but will look it up, thankyou. Totally agree about each number being a person, it is devastating.

It seems such an important distinction to me at this stage, the two different sets of data paint quite different pictures.

OP’s posts: |
donquixotedelamancha Sat 25-Apr-20 23:53:54

If you look at the graph by the date deaths are recorded, you can see a clear downward trend.

Because the more recent days will not have all their deaths collated yet, so it will always look like a downward trend for the last few days.

If you want to tell whether we are at or over the curve you need to take a rolling average of new cases (or deaths) and plot it against the total cases (or deaths) to see the rate of change as a proportion of the total.

What this tells you is:

1. We are just past the peak.
2. New cases aren't going down much so deaths will continue to be high for the next 2-3 weeks at least.
3. Our rate of decline is a bit slower than comparable countries. It looks like our curve might be very gradual.

Inkpaperstars Sun 26-Apr-20 00:24:03

I thought this graph had been shown on the daily briefing. They certainly did show figures that have been reported in the past 24 hours versus when they occurred as above. Certainly the graph has been widely shared.

BigChocFrenzy Sun 26-Apr-20 00:28:28

"Because the more recent days will not have all their deaths collated yet, so it will always look like a downward trend for the last few days."

Yep, some of "today's" deaths happened 4 weeks ago !
but it can take a lot of time for a death to go through the system of certification, notification, collation

BigChocFrenzy Sun 26-Apr-20 00:31:35

"They just want to make it look bad."

No, the headline numbers don't even include the care home deaths, which would bring the UK up to about 40,000 deaths in total

Belgium, France, Germany and iirc Ireland include care homes in their death totals

So the UK actually looks better than it is

Teapotdespot Sun 26-Apr-20 00:36:26

This isn’t accurate. Deaths can take over a week to be reported so of course the most recent days look like there’s a reduction in deaths, it’s because they’re not all reported yet.

I’m a stats geek by trade, reporting deaths from epidemics is notoriously difficult to plot on a rolling daily total for the reason above. There’s absolutely no conspiracy here but if you really want to get juiced up about something, focus on why the government have excluded deaths in the community to try and make our figures look better than the rest of the world...

donquixotedelamancha Sun 26-Apr-20 00:39:23

No, the headline numbers don't even include the care home deaths, which would bring the UK up to about 40,000 deaths in total

I'd be very interested to know if you have a source for such a huge number? That would suggest just under 7% of all care home occupants have died in about a month.

Greysparkles Sun 26-Apr-20 00:44:16

Because it doesn't fit the narrative

They want us to be scared. When they say 800 deaths today it sounds alot more scary than the actual truth.

What better way to control the population into taking lockdown seriously than through fear.

Even though personally I don't think the majority of the population needs to be scared at all as the fatality % of this virus in the healthy population is very, very low.

JasonPollack Sun 26-Apr-20 00:44:30

I've seen many graphs which look like that.

I'm not surprised that you haven't though if you're only listening to the radio. Not many graphs on the radio are there.

buttermilkwaffles Sun 26-Apr-20 00:48:31

@donquixotedelamancha

Probably the FT?
mobile.twitter.com/ChrisGiles_/status/1252841436317315072

Inkpaperstars Sun 26-Apr-20 00:52:50

What better way to control the population into taking lockdown seriously than through fear.

If they were trying to scare people they could very very easily give us many more of the gory details and as BigChoc says include more emphasis on deaths outside hospital. They are not trying to make things look bad.

What the population need if they are struggling to take lockdown seriously is not fear, it's logic and the willingness to engage with information for a short while. Unfortunately that seems almost entirely missing in many people.

donquixotedelamancha Sun 26-Apr-20 00:56:21

Thanks buttermilk. The big problem with the FT analysis is it assumes that all the extra deaths are due to Covid 19.

Still they are likely to be in the right ballpark which is fucking terrifying.

donquixotedelamancha Sun 26-Apr-20 00:58:15

Because it doesn't fit the narrative. They want us to be scared. When they say 800 deaths today it sounds alot more scary than the actual truth. What better way to control the population into taking lockdown seriously than through fear.

Are you on glue?

HeadPotato Sun 26-Apr-20 01:03:44

This article mentions a better way of measuring deaths

dannyboycorona.blogspot.com/2020/04/just-numbers-2-these-people-raised-us.html?m=1

lockedown Sun 26-Apr-20 01:25:19

The government doesn't want the numbers to look bad otherwise they would include the number of deaths outside few hospitals as well ( like few countries are doing already).
The number seems to be decreasing rapidly because today's deaths will also be reported much later and so on. Like you wouldn't have concluded by looking at this chart on 10th april that 8th april was the peak.
We just crossed 20000 hospital deaths. It is real.

dementedpixie Sun 26-Apr-20 08:22:35

They report on deaths outside hospitals on a Tuesday in England and wednesday in Scotland (or the other way round, not sure). They are reported but only on a weekly basis

Join the discussion

Registering is free, quick, and means you can join in the discussion, watch threads, get discounts, win prizes and lots more.

Get started »