Data and analysis thread, started 12 November

[NoGoodPunsLeft]

Previous thread here:

Data and Analysis Thread, started Oct 29 www.mumsnet.com/Talk/coronavirus/4064113-Data-and-Analysis-Thread-started-Oct-29

Regular lurker but I frequent poster, didn't want to lose the threads.

@IceCreamAndCandyfloss interesting, I just tried to get a test to see, initially I got the error message that there were none available, so I checked back through my answers and I had put my date of birth instead of date symptoms started . When I corrected this error I had an option of 150 walk in appointments or home test kits. I’m in London, so there doesn’t seem to be a shortage here,so may be localised

18213 cases
696 deaths

@GingerLemonTea

18213 cases 696 deaths

Where did you get that please? Dashboard isn't updating for me.

Ah, it has now.

I’m getting the same figures on dashboard as @HoldingTight

Sorry meant to tag @GingerLemonTea

@wintertravel1980

Tuesday numbers are generally low for cases (and high for deaths). Unexpected spikes in positive tests tend to happened on Wednesdays and Thursdays (when Monday tests get processed).

This time official data is consistent with Zoe (Tim Spector believes R in the UK dropped to 1 on Oct 31). There may be some backlog of cases due to the weekend but I would not expect it to change the downward trend.

Quoting this from yesterday predicting a Wednesday jump in cases

Well 15267 cases added in England; 11514 with spec date for Monday. Takes Monday to 12099 so far as opposed to 15412 for last Monday this time last week.

Those questionnaires linked upthread are very poorly designed. Whoever gave the example “primary teacher” didn’t think it through. They must know people often don’t read things properly, so someone who actually notices “primary teacher” might put “secondary teacher” or “FE teacher”, whereas someone who didn’t read it properly (or a poorly trained second party administering the questionnaire) might just put “teacher”. The ONS? I’ve marked down A level students for similar.

@borntobequiet

Those questionnaires linked upthread are very poorly designed. Whoever gave the example “primary teacher” didn’t think it through. They must know people often don’t read things properly, so someone who actually notices “primary teacher” might put “secondary teacher” or “FE teacher”, whereas someone who didn’t read it properly (or a poorly trained second party administering the questionnaire) might just put “teacher”. The ONS? I’ve marked down A level students for similar.

And also it depends on whether people actually think about what they might want it for. I asked DH and he said if he didn't think much about it he'd write 'history teacher'. Which would be a useless answer because it implies secondary but not clearly enough, and what kind of school he teaches in is more relevant to infection risk than his subject. But it's a similar level of detail to 'primary teacher' and he said he wouldn't necessarily have thought about what the most useful thing to put would be, just what he most often describes himself as

Exactly to last two posts.

If I had a client commissioning that research they’d likely be pretty annoyed and say it wasn’t robust as poor design. Especially against conclusions drawn.

Am I just being cynical or rather: how do you guys interpret the data? Just looking at England, pillar2

In the past 3 days testing has fallen to really low levels and thus cases are falling.
I think although lockdown(ish) might be having an effect, it's more along the lines of artificially pushing down case numbers by testing less.

1. I really can't tell any more if I'm looking for things that are not there... this gov has screwed up so many times and lied openly that I do think it's an option.
2. Given the first point it still stands that gov might directly have nothing to do* with how many ppl go and get tested.
3. It would be good to have an explanation for the 24th. we'll see in a few days if that was the glitch or today, but it'll be interesting.
4. another thing that caught my eye is although they halved the number of tests from let's say 18th, the number of cases went down, but not halved. To me this indicates a possible still rising trend?

*albeit through their messaging they do have an impact obviously.

Testing / cases
today testing numbers are not available yet for England

24-11-2020 173,219 / 9,854
23-11-2020 105,740 / 13,329
22-11-2020 153,430 / 16,668
21-11-2020 226,566 / 17,615
20-11-2020 248,122 / 17,845
19-11-2020 238,736 / 20,291
18-11-2020 211,629 / 17,189

@amicissimma

Take the headline the other day that supermarkets were the main transmitter of coronavirus, followed by secondary schools then primary schools... add secondary and primary schools together and it turns out that schools are actually the main source of coronavirus.

This was NOT what the report said. It said that the place most people who tested positive had been in the days before their test was the supermarket followed by school. As there were very few other places anyone could go, most people would have reported a supermarket and/or school as the place they'd been in the previous few days, including those who didn't test positive.

Please can we put this misinformation to rest.

But presumably quite a large number of people would have gone to the supermarket -- there are 56+ million people in England. And only 8.89 million school pupils (explore-education-statistics.service.gov.uk/find-statistics/school-pupils-and-their-characteristics)

So I'm surprised that the percentage who've reported schools is so high.

@borntobequiet

Those questionnaires linked upthread are very poorly designed. Whoever gave the example “primary teacher” didn’t think it through. They must know people often don’t read things properly, so someone who actually notices “primary teacher” might put “secondary teacher” or “FE teacher”, whereas someone who didn’t read it properly (or a poorly trained second party administering the questionnaire) might just put “teacher”. The ONS? I’ve marked down A level students for similar.

I completely agree. Surely there should have been a drop-down menu for types of educational setting!

I think a lot of people are naive and think that this is poor design on the questions. I think it's "clever" design for the purpose of being able to use data to underpin the narrative.

Parallel to uni I worked briefly for ipsos and it was eye opening :(

It does seem awfully strange that a survey would be quite so poorly designed.

@TheSunIsStillShining

I think a lot of people are naive and think that this is poor design on the questions. I think it's "clever" design for the purpose of being able to use data to underpin the narrative.

Parallel to uni I worked briefly for ipsos and it was eye opening :(

I’m not sure about this! It’s clumsy rather than clever. I agree research can absolutely give results you want by design but why is the unknown teacher bar not included in result in any case?

A better design would be to not leave an open hanger like this. We can all see the glaring mistake.

@mother. Yes but those who are students (or teachers) go to school every day. As compared to one person per household once a week for most supermarket visits. (Or none at all if you use online deliveries) And that's assuming you check in with your app at the supermarket. (I don't tend to do so places I know I'm not staying long) Which is why we need some real comparison data.

I wonder if that section was a badly worded question but then included because it fitted a story. Now that certainly fits what I’ve seen research / client presentations. The whole section bugs me - including the conclusions.

I think the problem with the ONS survey is that it was not designed with the primary objective of finding out whenever teachers had higher rates than other occupations. It's key purpose was to track overall rates of infection in a standardized way that didn't vary depending on levels of testing in the wider community. They've tried to use the data to compare age groups, regions and other things if interest (such as occupations) but that didn't set it up with a single minded objective of seeing whether teachers happened to have higher rates than other groups. Hence the definitions and sample sizes are non- ideal.

Yes, good point about supermarket visits being less frequent, @Firefliess.

I don't think it can be to do with checking with with the app, as that isn't possible in schools (no QR codes, plus primary school children wouldn't be able to).

TheSunIsStillShining - I have been having the same thoughts over cases versus tests. This data is easy to manipulate and I don't trust that it is not happening. I do believe looking at the zoe data that cases are decreasing but not as quickly as we are being led to believe.
Are the pillar 2 testing numbers available on a local level?

It will be very interesting tomorrow to see which tiers we are all in.

I agree with both Marsha and Fire. Badly designed questionnaire, results appears to shore up convenient narrative about teachers not being at risk, government jumps on it and now every time an MP gets a letter worrying about infection rates in schools, the misleading stats are trotted out.

The DfE asked MN to put a statement on a schools thread! It, cites the ONS survey. So it really ahs become part of the propaganda machine.

Yes, Piggy - I had that in mind too! Maybe I’ll report them to themselves warning them not to quote dodgy stats courtesy of the DfE.