All of this to say, there isn't an easy way to meaningfully compare case counts. Ideally we could calibrate test positivity to the total number of infections, using serosurveillance data. Instead of absolute numbers, I find it more helpful to examine trends. 10/11

But we can dream up settings where it's not linear. Imagine you only test a few very sick people. They are 50% positive. We test more people, and they are also 50% positive because they are like the others. We have captured more infections, but positivity hasn't changed. 9/11

The adjustment looks like this. Whatever the test positivity was (10%, 50%), multiply the number of new cases by that (10X, 50X). This assumes that there is a linear relationship between test positivity and the amount of the epidemic we are missing. 8/11

Can we adjust the numbers to make them more comparable? @foxjust proposed a simple rule of thumb approach. I think it's helpful to flag that there were likely more new infections at the mid-April peak than now, but of course nothing is easy. 7/11

As we have greater capacity for testing, we can move towards the outer tiers, and the test positivity declines. Of course it also declines with how much the virus is circulating in the area. 6/11

Each tier has its own probability of being positive. Someone with severe respiratory illness could have COVID, but they could have something else (depends on how much COVID is circulating). Let's say hypothetically 50% are positive. I added some numbers to the figure. 5/11

We hear a lot about test positivity being an important metric to track, but I think it's helpful to remind ourselves why. When I think about testing, I tend to imagine dividing the population into tiers. The center is most likely to have COVID. The outer layers less so. 3/11

At the peak of NYC's epidemic, over 50% of tests came back positive. Compare that to Florida where positivity is still below 20%. 2/11

THINK LIKE AN EPIDEMIOLOGIST: There are more new confirmed cases each day in the US than at any time during the earlier April peak. But is it really meaningful to compare those numbers? How do epidemiologists decide when to sound the alarm? A thread. 1/11

This but for ALL papers, ALL arguments, ALL articles, ALL op-eds. This is the future I want. Author doesn't even need to generate them, though would be better if they did.Quote TweetJasmine Wang@j_asminewang · 29 Aug 2019it'd be excellent if all philosophy papers included a diagram illustrating the paper's argumentation structure in the abstract. it'd be even clearer if there were also credences (%) assigned over the nodes here. diagram by @JohnDanaher

#MentalHealthAwarenessWeek: our webinar will explore the mental health impacts of #COVID19 across the life course & share how councils are working with partners to support public mental health and wellbeing. Join us on Thursday → http://socsi.in/gtRTr

(1/n) A rule-based experiment of coupling a social decision-making model with an infectious disease model to explore mask wearing. A thread. (H/T @davidmanheim @vee3my) #epitwitter #MaskUp #COVID19

(8/n) You can run this model in the browser and experiment with different parameter regimes to get different behaviours and equilibria. It's a little CPU intensive, but it works.

(7/n) This is reminiscent of what we called "information cascades" where decisions are made based on a mix of public (copycat) and private (watching the news) information. This has a game theory flavour because m_xy encodes the altruism of wearing masks.

(6/n) But how does the first person decide to wear a mask and start the mask wearing cascade? Who will we copy?Perhaps those paying attention spontaneously wear and remove masks proportionally to the current danger, the number of infectious people.

(5/n) Following the crowd means, if I see people wearing masks, I'm likely to decide to wear a mask. This has much the same shape as an epidemic. An epidemic of masks. We can also do the same thing in the other direction: if people aren't wearing them, perhaps I won't either.

(4/n) How do we know who is wearing a mask? This is where it gets interesting. Suppose there are two mechanisms at play: following the crowd, and watching the news.

(3/n) If nobody wears a mask, all the virus gets through. Only the susceptible one does, helps a little. If the infectious one does, it helps a lot. If both do, it's much better. There aren't many empirical studies about "source control" so the numbers used are arbitrary.

(2/n) The infectious-disease part is what one might expect. Transmission is attenuated depending if either or both susceptible and infectious individuals are wearing a mask. Otherwise it's just your standard SEIR-style arrangement.

This is an example from our paper "Scaling up epidemiological modelling with rue-based models". Rule-based modelling turns out to be a very clear way to express models with rich interactions and simulate them efficiently. The pre-print is here: https://arxiv.org/abs/2006.12077

This is a governance failure, not an inevitability of the disease.

Current mood of infectious disease epidemiologists in the US? Stressed.

We really do need to know. The Johnson Government really need to tell us explicitly. Because the lives of nearly 50,000 people a year (based on current death rates) depend upon it.

Is it to accept the present stalled rate of infection and death? Is it herd immunity by default? Or is it (as it seems) a series of ad hoc measures driven by olitical experiency and whoever is pessing the Government hardest?

A very simple question to start the week. Scotland now has a clear strategy to drive towards elimination of the virus. But what is the strategy in England? We have a series of policy changes - open the pubs, end shielding - but we have no clear overall statement of strategy.

The effective reproduction number Re is now above 1 in #Switzerland: 1.28 (95% 1.06-1.53). We *urgently* need an in-depth understanding of transmission dynamics, the effectiveness of contact tracing etc. And #MaskUp

8.8 In summary, are we still in control? Currently yes, but just. This whack-a-mole is dragging, and our concentration and arms are tiring. All are putting the effort, but I feel that holding on will depend on what we do individually. As they say, salvation lies within.

7.8 Meanwhile, the financial turmoil and civil unrest continue. When people are desperate or angry, they become undisciplined, risking their safety. Even stable societies are showing “virus fatigue”, and want their social life back. Balancing ‘normal’ and ‘safe’ is proving hard.

6.8 Nearly 2% of repatriated passengers tested +ve. This rate when applied to the expected 2000 daily travelers results in 40 new +ve cases daily. The actual number may be less. The situation with C-19 in the US, Gulf, Africa, and recently Europe, does not support this optimism.

5.8 The surge in clusters with large number of contacts is stretching our resources to trace and test. The locations are also difficult to isolate. This may lead to delays in containment, which may further lead to spread. This vicious cycle, seen elsewhere, has to be avoided.

4.8 New clusters were reported in Beirut, its suburbs, in a hospital, and in refugee camps. Densely populated locations allow more contacts per individual facilitating spread. Worryingly, several of theses clusters remain without a clear index case.

3.8 Some argue that since our capacity can absorb this expected rise in hospitalized cases, we should not be alarmed. The transfer of 3 cases this week to RHUH from #Covid19 ready hospitals due to lack of certain services raises a few questions regarding our preparedness.

2.8 The number of new +ve local cases has increased. The last 7 days have witnessed almost tripling in the number of cases as compared to the previous 7 days (from 47 to 143 cases). This surge usually results in an increase in hospitalizations, though it is too early to tell.

1.8 Is our #Covid19 situation in getting out of control? Several recent observations can help answer this question:

SciBeh now has a video describing our initative! watch, retweet....

Sean Hannity on May 20: "Florida got it right, Texas got it right ... the mob and the media, by the way, you own Governor DeSantis a huge apology"

RT @ScottGottliebMD: See how quickly you can find Sweden on this map....

in other words: overestimation of familiar risks makes COVID-19 seems much less of an outlier, and much more like background 'everyday risk' than it is which would explain why so many otherwise smart (but non JDM) people I see keep churning out that phrase

But (=lightbulb moment) these two things are entirely compatible, because that over-estimation means that the many non-COVID daily risks of death people are faced with, are assumed to be much more prevalent (driving, flying etc)- and, "experience" suggests we can live with those

(to which I've replied, like a broken record, and you think that's LOW"?) ..this has sat uneasily with the one thing we all think we know, namely that people are supposed to be unduly freaked by really low probability events, because they robustly overestimated low probabilities

Fellow behavioural scientists, I had a lightbulb moment yesterday. I suspect I might have been unusually slow here, and you all clocked this months ago, but thought I'd share nevertheless: I've lost count of how often I've been told "You only have a 1/100 chance of dying"

Coming out soon in @terpolv! I'm very excited about the release of this study. Years in the making.

Unfortunately, uncertainties in variation in contact rates and susceptibility/infectiousness will prevent us from ever having precise value for HIT for 1 pop & it will vary among populations (as does R0). The take home message is want to get to HIT by vaccination NOT infection.

Together papers suggest herd immunity thresh is <1-1/R0 and epidemic overshoot <f=1-exp(-(f*R0)), but by how much is still unknown. Ugh. @mlipsitch smartly gave a wide range 40-70% (20-60% in later interviews) for fraction infected f w/out interventions.

Serological data from NYC suggest that either heterogeneity estimate is too large or R0 is much larger, or lockdown did nothing to stop outbreak as seroprevalence is already at overhsoot epidemic size f (~20% - I added horiz dashed line).

I'd like to see uncertainty around those estimates of heterogeneity & some are very indirect estimates, but range of values is much better than arbitrarily chosen numbers and suggest heterogeneity is substantial. But ...

@mgmgomes1 paper is much better in providing empirical estimates of heterogeneity & impacts on HIT, f based on data (colored lines in figure).

What I find puzzling is that this paper is in Science. It's a extremely simple modeling exercise, with arbitrarily chosen parameter values. It isn't as general as @mgmgomes1 paper cited above, and it's not clear it has any new empirical data. Many older papers say same thing.

The conclusion in the abstract is appropriately worded in making it clear that the paper is simply an illustration of a concept, not an actual estimate of the most likely value.

The take home of the science paper is that pops partially structured by age (i.e. NOT well-mixed) and w/ arbitrary differential "activity" (contact rates) w/in ages can reduce HIT for R0=2.5 from 0.6 to 0.43.

The strange part is that it's not clear what source their age-mixing data comes from. The citation they give is NOT a published paper - it's a link to a page w/ Matlab files. It's weird because there are several age-mixing datasets out there. @sbfnk can probably provide best ones

Science paper (link above) also shows how reasonable variation in contact rates among individuals reduces HIT. It considers differential mixing among age groups and hypothetical variation among individuals w/in each age group having more or fewer contacts.

However, see paper by @mgmgomes1 that is even better (b/c more general, better data) than Science paper cited above. Fig shows how variation susc, contacts affects HIT, f, for R0=3; so HIT=0.66, f = 0.94.

Positive correlations among these traits - if individuals w/ more contacts are more susceptible and infectious - can make for very rapid changes in dynamics. Sadly not much empirical data on these correlations & limited data on predictors of variation on susc. & infect.

1) If some individuals have few contacts then they are unlikely to be infected, lowering HIT. (note: variation in contacts can lead to faster early spread). 2) If some aren't susceptible to infection that reduces HIT. 3)If some people don't transmit that also reduces HIT.

Human populations are NOT WELL-MIXED & HOMOGENOUS, so calcs give likely upper bound. Q: So what is herd immunity threshold (HIT) for real pops? A: *Usually* lower & it depends on variation among individuals in 3 things: 1) contact rates; 2) susceptibility; 3) infectiousness.

These simple calculations & flu data are what led @mlipsitch to make early widely cited rough back of the envelope calculations that 40-70% of pop would get infected without interventions. Most articles cited the 70% *BUT* @mlipsitch knew that...

In idealized population, early vaccination to HIT will prevent outbreak b/c Rt=R0*S/N so Rt<1. However, an epidemic will overshoot HIT & infect larger fraction f of pop where f=1-exp(-(f*R0)). For R0=2.5 (reasonable for COVID19 in some pops), HIT=0.6, and f = 0.89.

Background: If we know reproductive number R0 for pathogen in population, herd immunity threshold (HIT) for WELL-MIXED HOMOGENOUS population is simply 1-1/R0. (Note R0 varies w/ behavior, host traits, environment so differs over time&space).

What is #herdimmunity threshold for #COVID19? A topic of much discussion due to large uncertainty & huge consequences for long term impact. New paper illustrates 1 key aspect that can reduce it substantially (spoiler: we still don't know value). Thread.

Based on current rate of daily case growth (which I est. at 3.4%), I am predicting that we will have 100k cases PER DAY on or around July 24 if we don’t do something drastic AND soon! Even with measures taken, we will almost certainly still achieve that level, though in August.

Democrats & Republicans have seemingly taken a divided stance on how to handle the #COVID19 pandemic, & their Twitter accounts might provide the best evidence to date. Watch the latest video from our #coronavirus series on research from @ScienceAdvances: https://fcld.ly/kp8s9er

#sfhphealthpolicycouncil Help protect at-risk populations from extreme heat during the COVID-19 Pandemic

Have courage. Be kind. Keep trying. We can do this — we have to do this — for them. END

4) In the end I think of the lesson from the bunnies. We have the keys to make this a better world. Those keys are logic, science, reason, and data. But the more important ones are: kindness, communication, open hearts and courage.

… but I can’t look at my students, or my kids, or the young people I mentor, and feel anything but an obligation to do what I can for them. No matter how tired or cynical I feel.

3) The next generation needs us. They are amazing and hopeful and inspirational and trying so hard to make something of the shit sandwich they’ve been given. But they can’t do it alone. I’m old and cynical and exhausted…

2) The world seems to be getting harder. There’s no denying that. Hard things suck, but they can be cleansing too. I have had *such* a hard time lately, and it’s still hard. But it’s made me better: I know who I am, and I know how to be that person.

1) There IS no science/art divide. Both are so important. Math, coding, informed data analysis, proper scientific method — these are keys to understanding. But art — stories, like the bunnies — that is what gives it all heart and meaning.

… I like to think *we* learned to rely on each other and to do the same. Even online, I felt like we were all pulling each other through this difficult time. And, looking back, I see a lot of lessons for myself that I want to remember.

And through all these terrible months, these students (and the bunny story) kept me going. Just as the bunnies learned to rely on each other and use data and science and reasoning to solve their problems…

… careers in data science, in analysis, in psychology, in all kinds of social science. I’ve told them I’m so proud of them but I don’t think they grasp HOW proud of them I am and how completely impressive they all are.

And in this mastery, they’ve found confidence. Some have realised that they actually DO like math and coding. Some have realised that they can do really hard things. Many now have the foundations for all sorts of careers…

THEY STEPPED UP, ya’ll. These students, who three months ago knew no stats and no coding, somehow in the middle of a pandemic managed to master this stuff at a level that will open major doors for them in the future.

… for an inability to differentiate students. I had to make the exam difficult for this reason, and I was afraid they’d hate me or not be able to do it. We’re still in the middle of marking it but neither of those things seem to have happened.

Some students struggled, and I have so much respect for them, because they KEPT TRYING. Most of them ended up getting it, or at least getting most of it. On the first assignment, which was not easy, the grades were so good I worried about getting in trouble…

Some students went so far beyond expectations: finding other datasets on the web and playing with them for fun; asking highly technical questions about the central limit theorem and the foundations of statistics; asking for more math, more resources.

And yet, OMG, they persevered. They believed me (or at least tried to believe me) when I told them they could do it. They didn’t let their fear stop them from learning more about R and stats in a few months than many people had told me they’d be capable of.

Plus they were taking my subject — a subject most of them feared and thought they’d loathe. A subject that is really quite difficult.

As hard as this semester was for me, it was also really hard for my students, if not harder. Many didn’t know what country they’d be living in. Many couldn’t support themselves anymore, or had terrible living situations, or no childcare.

I’ll tell you. The hope is the students. The hope is this next generation we have coming up. They are spectacular, y’all, and we fucking OWE it to them to step up and fix this shitshow of a world we’re leaving to them.

Where’s the hope? You may be asking at this point. So far this seems terrible. Where’s the grace?

My subject was better adapted to online learning than many — we already did flipped classroom, for instance — but it was still a huge source of stress and uncertainty and extra work for myself, the tutors, and the students.

Then Covid hit and it went from difficult to feeling at times impossible. We have two small children, ages 4 and 7, and while my partner has been fantastic at taking the lion’s share of the unexpected child-rearing, it still hit me too.

(And of course in addition to this subject I also had a lab to run, honours/PhD/etc students to support, grants to administer and write, and colleagues who depended on me to work on joint projects.)

The bunny story added to the workload even more but I couldn’t not do it. Anyway, all of that is to say that even before Covid, I knew it was going to be a tough semester: SO much work on top of the substantial personal stuff I’ve been dealing with.

.. and over the semester they would learn to use data and come together to solve their problems and find a way forward. I was afraid the students would find it a bit cringey, but at the same time I thought it sounded so fun I had to give it a try.

While planning the subject I came up with the ridiculous idea of wrapping the entire semester around an ongoing story centred around my children’s stuffed animals. It would be a tale of bunnies facing food shortages, terror of outsiders, fear of the unknown…

Still, I asked for this. I love teaching SO much, and I love teaching coding and math more than any other kind. Coding and math are my happy place — meditative, soothing — and I love the challenge of getting people who fear and loathe it to see some of its beauty.

In March I started teaching an #rstats class at the intro level to almost 700 psych undergrads. It was my first time teaching it, which meant spending months on an insane sprint creating 25 lectures, weekly tutorials, assessments, and answering zillions of emails.

I’ve been having a difficult time lately — partly because of [insert frantic gesturing at the state of the world], partly personal — but one thing has been a real bright light for me in the last few months. I think it has some broader lessons that might give some hope, so THREAD

“It wasn’t diverse enough, maybe it wasn’t nimble enough” Sir Paul Nurse tells @Peston that SAGE needs to be truly multidisciplinary to perform its duties. #Peston

In the midst of a global pandemic, @BU_Tweets has just threatened to pull the medical insurance of any graduate student unwilling to teach in the classroom this fall.

Science is not our missing ingredient in beating this virus. Empathy is. /end

As the death rate drops & people feel safer, lower risk people will be even more cavalier about getting back to their lives. 24/

I don’t know Joe Rogan but do people in our country really care what he thinks about them? While local officials & scientists get death threats for wearing masks? Get asked to falsify data & under test? What is wrong with us? 23/

Where did they teach in Sunday school we can be negligent of others when we feel safe? Where is pious Vice President Pence? Where are evangelical leaders? 22/

If I told you there was a force that preyed on the old, the sick & people of color, like my friend @byron_auguste you might refer to it as a Nazi disease. 21/

What is the matter with us? Why do we need to personally be at risk to care about risk to others? 20/

But simple things like wearing a mask, not attending large events like casinos & churches, not hanging out in bars protect everyone & cut down on transmission. The 9 of the 10 cities with the highest growth in transmission are in places with low reported mask use. 19/

The likelihood of catching & the severity of the virus goes up as prolonged exposure & the number of people you are exposed to does. Running into a grocery store without a mask is a different from working there.18/

If you have autism or a developmental disability, you are more likely to get COVID AND if you get it, much more likely to die from it. Same with high blood pressure, obesity & other illnesses. And age— fatality climbs from 1-20% between 50 & 80. 17/

We all likely know by now that Black Americans are 2x likely to die of COVID-19 as white people. But if you’re between 35-44, Black Americans are *9x* more likely to die of COVID-19 as white people. 15/

It is becoming clearer & clearer that we have a situation much more like the second scenario. Young people feel immune & many are symptom free White people have lower mortality Well off people can isolate But they can all easily spread COVID-19 to people at higher risk. 14/

Let’s not pull punches here. In other countries they faced this question & society has rallied. All over Asia, Europe & Oceana, people have worn masks, locked down, sacrificed income & stayed away from bars. It wasn’t easy & they might have to again but it worked. 13/

Or would they say “ let me live my mask free life to the fullest, the high risk people can take care of themselves & isolate.” Can people working essential jobs perfectly isolate? Should they have to if a month with a mask for everyone is an alternative? 12/

To keep it simple, would lower risk people wear a $1 mask to dramatically cut down on transmission knowing they were going to come into contact with higher risk people or their family members even better even if they weren’t worried for themselves? 11/

Here is the question— if the transmission rate was very high, do we live in a society where young, white healthy people would be just as cautious as Black & Brown people, older people & sicker people? 10/81871.5K

What if young, white, healthy people had a 25% chance of catching DIVOC & a 1% chance of dying (.25% overall). And people of color, sicker & older people had a 40% chance of catching it & a 20% chance (8% overall) of dying? 9/

Let’s say there was a highly infectious disease called DIVOC- 91. If all people had an equal 10% chance of catching DIVOC & a 20% chance of dying (or 2% overall) if they got it, everyone would be careful. It’s natural. We want to live. But what DIVOC were different? 8/

Scientific progress was made possible by all of our actions in March & April. We stayed home, we distanced. To everyone who did, you saved lives. Now there is an important question about that behavior. Was it done to protect ourselves or also to protect the people around us? 7/

In a society where we look out for one another thie scientific & social adaptation be good. But societies where many people largely look out for themselves will have a different outcome. 6/

When the virus was in Wuhan, being on a ventilator was a death sentence. 80% of people on ventilators died. Today that number is lower & getting lower still. With new therapies, that 80% could decline to 20-40%. 5/

Let’s start with the scientists. Science is attacking 4 major problems right now: -the virus -3 main complications (clotting, oxygen deprivation, immune system over-reaction) 4/

As time goes on and the virus spreads, 2 things happen: -our own behavior adapts to what we know -scientists adapt to what we learn 3/

The pattern in the US, in all likelihood, will face a pattern of: -cases going up -death rates going down -much more unequal outcomes 2/

COVID Update June 22: In a lot of ways, COVID-19 is forcing us to answer the question— what kind of society are we? Particulatly as the healthy, white & well off find ways to protect themselves will we look out for each other? 1/

A government client is seeking to review and build on the recent relevant behavioural research. Please visit this document to learn more about each of these areas and contribute your knowledge:

Are you, or is anyone you know, researching how COVID-19 has affected behaviour and behavioural drivers in Victoria and Australia, in particular behaviours that related to topics such as ‘active transport’, ‘service provision’, ‘working from home’ and ‘car usage’?

Reality check: Germany (pop 83million) running 200-300 cases per day and sudden outbreak is 1000 cases from a factory. UK (pop 66million) running 1200-1500 cases per day. Like an equivalent big outbreak everyday in the UK.

BAME & low-paid groups at particular risk as disproportionate number unable to work at home & exposed to greater risk under new rules expected to be announced Tues. Independent SAGE says govt must release underlying evidence so public & businesses can make own risk assessments

Independent SAGE agrees with government’s own SAGE group that current levels of transmission are too high. Reducing rules from 2m to 1m will effectively end all social distancing in UK. It is too soon to do so.

NEW: Independent SAGE has evaluated the scientific evidence on social distancing & concludes it is not safe to reduce it from 2m to 1m indoors as government proposes

RT @subatomicdoc: Amazing what a difference four weeks makes since Memorial Day Weekend in #COVID19 New cases/day by U.S. region, adjusted…

BAME & low-paid groups at particular risk as disproportionate number unable to work at home & exposed to greater risk under new rules expected to be announced Tues. Independent SAGE says govt must release underlying evidence so public & businesses can make own risk assessments

Independent SAGE agrees with government’s own SAGE group that current levels of transmission are too high. Reducing rules from 2m to 1m will effectively end all social distancing in UK. It is too soon to do so.

NEW: Independent SAGE has evaluated the scientific evidence on social distancing & concludes it is not safe to reduce it from 2m to 1m indoors as government proposes

UNESCO has produced a report on the  COVID-19 Educational Disruption and Response. The main points are:‘Most governments around the world have temporarily closed educational institutions in an attempt to contain the spread of the COVID-19 pandemic.’‘These nationwide closures are impacting almost 70% of the world’s student population.’ That’s 1,214,075,186 affected learners.More than 190 countries have closed schools for over two months  – 90% of the world’s student population. School closures occurred rapidly, however, when it comes to their reopening, many countries are undecided on when, and how, and with a considerable degree of uncertainty on the way forward. According to UNESCO: ‘100 countries have not yet announced a date for schools to reopen, 65 have plans for partial or full reopening, while 32 will end the academic year online.’<img class="size-full wp-image-17649 aligncenter" src="https://www.cebm.net/wp-content/uploads/2020/05/Schools.jpg" alt="" width="590" height="436" srcset="https://www.cebm.net/wp-content/uploads/2020/05/Schools.jpg 590w, https://www.cebm.net/wp-content/uploads/2020/05/Schools-300x222.jpg 300w" sizes="(max-width: 590px) 100vw, 590px" />In a pandemic, the proportion of deaths among the young should increase, but this has not been the case.  A review of 72,314 cases in China showed less than 1% were in children younger than 10. Out of 16,749 hospital admissions in the UK, only 239 patients (2.0%) were under 18 years and 139 patients (1.1%) were under 5 years old. In Italy, three deaths have been recorded in the age group 0- to 19 years. In under 45-year olds, ONS data in England and Wales reveals that 384 (1.2%) deaths have occurred out of 33,365 COVID cases with   only two deaths in under 14-year-olds.From March to mid-April this year, nine students and nine staff from 15 New South Wales Schools in Australia had confirmed COVID-19.  735 students and 128 staff were in close contacts – no teacher or staff contracted COVID-19 and only one primary and one high school child may have contracted COVID-19.A French study that identified secondary cases linked to the index case reported  that one symptomatic child, visited three different schools but did not transmit the disease despite close interactions.The risks of COVID transmission in children are low. Going forward, we will badly need their future knowledge. Children’s education and their wellbeing is –  and should be – a priority. Prolonged lockdown of schools penalises an entire global cohort. It incentivises excessive reliance on electronic means of communication and a sedentary lifestyle.  

Amazing what a difference four weeks makes since Memorial Day Weekend in #COVID19 New cases/day by U.S. region, adjusted for population

How to bootstrap confidence intervals of regression coefficients in JASP. #stats #openSource

But if we look at fraction of tests that are positive we see a very strong signal in which ages 0 to 17 are nearly flat through time while 50 and over show a strong decrease in test positivity from April to June going from over 25% to 4-5% who test positive. 7/10

t's difficult to quantify this, but it looks like infections may have declined more in vulnerable older individuals. This is possibly contributing to declining daily deaths. Though it's also possible that improvements to clinical care have lead to better patient outcomes. 8/10

Here, I downloaded COVIDView data and slightly prettified their plots. If we look at testing in commercial labs, we see that across confirmed cases, there are slightly more cases in the 0 to 49 year demographic than there were in April. 6/10

In the past 7 weeks, daily confirmed cases have decreased to ~80% of their April 30 value, while daily deaths have decreased to ~35% of their April 30 value. Figure from @nytimes. 2/10

A small follow up to the "long plateau" assessment of the #COVID19 epidemic. When I tweeted this on April 30, we had ~30k daily confirmed cases and ~2000 daily deaths. This last week, we had ~24k daily confirmed cases and ~700 daily deaths. 1/10

Yes, the distinction between the point I was trying to make and yours is subtle (and I don't think I distinguished them very well). I think under any conditions a 1m rule is likely to cause people to act at `normal' distances, possibly with the exception of with `close people'. However the wider signalling effect that you are talking about could possibly be mitigated by a replacement signal. For example, suppose people had to wear a mask when interacting in person with someone outside their household. The presence of the mask could then act as the signal for the wider `not normal' behavioural expectations. Possibly the mask could be equally as effective for that wider self-signalling but I am not convinced it would stop the distance normalisation i.e. the distancing rules are particularly and uniquely effective for enforcing `not normal' personal distancing.

good point about the study, and you are, of course, right about the different distances (starnger/acquaintance etc)- my concern really was about the new perception of "normal" as you describe (better than I did) in your fourth point!

A few comments.First, it's worth noting that the methodology of that social distancing paper was to ask people to assess based on a picture of two people what would be the right distance for a stranger, an acquaintance, and a close person. I think it would be reasonable to be cautious of the accuracy of people's perception of this distance from a picture compared to what they actually maintain.Second, the finding for UK (though the authors refer to it as England in one chart and UK in another) was that stranger distance was approx 1m, acquaintance distance approx 80cm, and close person approx 65cm. It seems reasonable to say that many contacts will be with acquaintances (e.g. work colleagues, regulars at the coffee shop / common room etc.), and with `close people' outside of people's own household (relatives, close friends). As others have stated it seems unlikely that people would keep to 1m compared to 80cm, or perhaps even 65cm. The Lancet paper suggests a doubling of infection risk for every metre under 3m. This translates to a ~15% increased risk of infection going from 1m->80cm and ~30% increased infection risk going from 1m->65cm. But these infection risks are per contact. An increased risk of these proportions in the risk of infection from a single contact results in a higher increased risk of infection of an individual (due to the multiple contacts they are likely to have), and would be expected to produce a still higher increase in total number of infections as the probability of contacts involving infectious people consequently increases. My point basically being that the population increase in infections is far from a linear relation with personal distance.Third, I wonder how much these norms might vary across the country.Fourth, and this is something that the Behavioural Scientists can probably comment on (it's kind of implicit in what other posters have said), I would guess the perception of something like personal distance is something on which we rely to some extent on a binary 'normal' / 'not-normal' rather than a gradable concept of distance (in metres for example). Putting the rule to 1m, puts it into the 'normal' category and we end up behaving as per that category (i.e. the 1m/80cm/65cm distances), whereas the instruction to keep 2m reminds people that they should be maintaining their distances in the 'not normal' category.

This animation shows the factors influencing the impact of #Covid_19 on #BAME communities

I hope the reader will forgive me if I say that my reaction to this is more likely to be spontaneous combustion than spontaneous applause. /10 /end

Aside: A friend who works for a large consumer products company once told me that when they make a 10,000-litre batch of their "herbal extract" shampoo, they add 10 millilitres of actual herbal extract to the vat. That's what this feels like. Semi-homeopathic open science. /9

So the article got the "preregistered" badge (arguably the journal's problem), and the authors themselves used the keyword "preregistered", on the basis of exactly one exploratory analysis, suggested by the reviewers, and written up in the supplement as having unclear effects. /8

Here's the preregistration. It was "preregistered" on June 3, 2019. That's 130 days ***after the manuscript was submitted***. It describes a preregistered analysis of... the possible effect of adding an extra explanatory variable to the models that are already in the paper. /7

Still, there must be plenty of other examples, right? After all, "preregistered" is a keyword. Let's find the next occurrence of "prereg". Oh. It's in the coda. Still, we get a link to this doubtless extensive preregistration. I might start getting excited again. /6

The next three mentions of "prereg" are in the Methods section, where for some reasons the authors seem to have become somewhat lukewarm about preregistration. ¯\_(ツ)_/¯ /5

(By the way, that image also contains the date on whiuch the manuscript was received: January 25, 2019. Keep that in your mind.) /4

Let's look for the results of any preregistered analyses in the article. The strings "prereg" and "pre-reg" find only a few hits. Here's the first: a keyword! Strong start there. This article is clearly going to be a feast of open science practices. /3

Clark et al. received Psychological Science's "preregistered" badge. Here are the criteria for awarding that badge, from the journal's web site. /2

Some interesting (to me) details about the retracted Clark et al. article. One thread per detail. This is thread #1: The preregistration. (There may only be one thread, depending on my motivation levels; it's a beautiful day here and the park beckons.) /1

Now we hope data we generated will help others studying evolution of virus and especially those developing vaccines. It's all publicly available, so if you're curious about a specific mutation to RBD, need a stabilizing mutation, or need to change a surface--go look it up! (14/n)

Allie Greaney jumped on board to help @tylernstarr do the experiments amazingly fast, and Sarah Hilton and @khdcrawford helped with code and validation. (13/n)

@tylernstarr had been optimizing approach for other proteins when #SARS_CoV2 hit. We had as colleagues @veeslerlab @coronalexington who had been building basic knowledge about coronavirus spikes for years. They helped advise @tylernstarr how to adapt his approach to RBD. (12/n)

To enable these deep mutational scanning titration assays to be applied to larger protein like RBD, we leveraged barcoding approach originally developed by @JShendure and subsequently extended to use PacBio by @lea_starita @kmatreyek @dougfowler42: https://nature.com/articles/nmeth.1416… (11/n)

Finally, how we were able to characterize so many mutations to this important protein? We leveraged great prior work. In 2016, @jbkinney, Walczak, & Mora showed yeast display & deep mutational scanning could measure dissociation constants at scale: https://elifesciences.org/articles/23156 (10/n)

Lots of other observations about natural mutations, structure-function, and sarbecovirus evolution in the paper itself, so take a look. (9/n)

Key finding from that comparison is that no antibodies have epitopes as constrained as actual ACE2. So epitope focusing can elicit more escape-resistant antibodies. And our data help point the way, as we show what mutations would be tolerated in resurfaced immunogens! (8/n)

We can examine mutational tolerance of epitopes of different antibodies, and compare it to mutational tolerance of actual ACE2 binding interface (here are some antibodies, see paper for more). (7/n)

We can map tolerance to mutations with respect to RBD folding / expression & ACE2 binding on structure (red = intolerant in images below). As expected, most selection for binding in ACE2 interface--but again, lots of mutations still tolerated (again see heat maps or paper) (6/n)

Some observations: While many mutations to RBD are deleterious, lots are tolerated--including a surprising number that increase ACE2 affinity (blue in heat map in prior Tweet). But no evidence that strong affinity enhancing mutations are being selected in human isolates. (5/n)

If you're interested in the evolution of the virus or are studying antibodies / vaccines, you can go to this interactive heatmap (https://jbloomlab.github.io/SARS-CoV-2-RBD_DMS/…) and simply look up the *experimentally measured* effect of any mutation. (4/n)

Before our study, the structure of the RBD protein was known but we didn't know how mutations affected it's function. Now we've measured how virtually all (3800 of 3819) amino-acid mutations affect binding to ACE2 and expression of the folded protein. (3/n)

The RBD (receptor binding domain) enables #SARSCoV2 to bind to human cells. Evolution to bind human ACE2 was key to the emergence of this virus. Now it's also key to mitigating the virus: the most potent antibodies bind to RBD, and most vaccine candidates contain RBD. (2/n)

We've experimentally measured how all amino-acid mutations to the #SARSCoV2 spike RBD affect ACE2 binding and expression of folded protein in a deep mutational scanning study led by @tylernstarr & Allie Greaney: https://biorxiv.org/content/10.1101/2020.06.17.157982v1… Why is this important? (1/n)

and last. government needs to use expertise on the ground. patient voices and organisations are really important here. yet seem scarcely used as a resource to help get this right. trust is not bought, is built. off to work

the second is false advertising. this should always been a red flag and needs to stop. no one in the NHS should trust any org/ provider who makes surreal claims about their stuff or who gets shirty when ask for evidence. yes, am talking about you @babylonhealth

the best tech has grown stuff from ground up and found out what is needed and wanted and what works. humility to know there's stuff you don't yet know. understanding that NHS staff can't be left fixing your botched rubbish.

Replying to @mgtmccartneythe first is belief that the private sector should get massive contracts for stuff that hasn't been done, and who work from top down. healthcare is really complicated. rests on trust and relationships. the gov approach starts off wrong.

profound disconnect between the government belief about what benefits technology will provide, and the later (after ££) evidence for benefit. tech in health=vital, but if not fit for purpose ends up wasting resources better spent elsewhere- time+money. what needs to change?

Numbers speak louder than words. What's the UK overall objective & which model to adopt? To be like Sweden (relaxed approach), like Germany (full control of the virus), or like NZ (eliminate it). Thanks for great website @Ian_Campbell_ https://iancampbell.co.uk/covid-19.php

RT @AmanpourCoPBS: "Spending in the U.S. as measured by credit cards has fallen by about 30 percent...which, just to be clear, is the bigge…

what impact it will have, I guess depends crucially on whether scientists can argue that 'best practice was followed' (see also Victor's comments).The problem is that, as you point out, we arguably don't yet have a consensus around code and data sharing, though there has been a huge push in that direction in recent years.I've personally been a bit lukewarm about some of that in the past, precisely because of the resource issue and the fact that often it seems to me to make more sense in terms of the field's resources to just run a study again.But I'd definitely argue that at the moment, the reverse is true: we need to spot errors more rapidly than in 'normal science'.

Interesting webinar opportunity from @LGAcomms

How We Reopen Safely

I am here to help.

Tomorrow at 1pm CEST I'll be doing a virtual talk for the Rotterdam R.I.O.T. Science Club (@rdam_riots) on using Twitter for science I'll be covering both the *why* and the *how* + I'll be leaving plenty of time for a Q&A session. Watch here: https://tinyurl.com/y7uoe4ls

Tomorrow at 1pm CEST I'll be doing a virtual talk for the Rotterdam R.I.O.T. Science Club (@rdam_riots) on using Twitter for science I'll be covering both the *why* and the *how* + I'll be leaving plenty of time for a Q&A session. Watch here: https://tinyurl.com/y7uoe4ls

A new discussion paper:"Psychology we argue, is unsuitable for making policy decisions. We offer a taxonomy that lets our science advance in Evidence Readiness Levels to be suitable for policy; we caution practitioners to take extreme care translating our findings to applications."

RT @AlistairHaimes: @d_spiegel ONS have now published 5-year averages by place of death, helps to put the numbers in context https://t.co/Y…

‘The thing I find the saddest is we knew the entire story of what’s unfolded in Britain in the last 4 months by January 31.’ Editor of @theLancet, @richardhorton1 answers @Piersmorgan's question on why he thinks the UK has handled the coronavirus pandemic so badly.

Here's Hillary Clinton posting the NPR coverage of the article. Again: no study, no data, just claims made by people who have not done very good research in this area in the past. Makes for a great headline, though.

I think I am going to have to do a more formal analysis of how this misinformation *about* misinformation makes the rounds and affects the conversation. I promise I will publish my data before making any press releases.

Just scrolling through the replies and quote tweets of the NPR and MIT Tech Review articles I have found hundreds of discussions and comments mostly supporting these articles, which are themselves unsupported information!

One irony is I *want* to believe that 1) there is a COVID bot problem on Twitter and 2) the bots have a measurable effect on the discourse. It would be a relief, in a way, to know that people are being manipulated into behaving poorly. But so far, no evidence either way.

I'm very mad today at news outlets who fall for the "trust us" line and will publish an article about the claims of scientists without linking to a published paper or at least a preview of the paper.

Extraordinary claims require extraordinary evidence, and this particular claim is accompanied by literally zero evidence. Instead, they are trading on the name of Carnegie Mellon University, saying "trust us, we're CMU scientists."

Here is my original thread responding to NPR's coverage of this press release without a study.

This story continues to be widely circulated today, this time though MIT Technology Review. Again: this is based on a press release with no published data or methodology, and the bot detection work by this research lab in the past is questionable. See my tweets above for why.

@BobbyAllyn I would happily talk to you about this issue and help set the record straight, you are welcome to look up my name on Google Scholar if you need to see the wide range of citations of my (non academic) work on bots

Also if you're interested in this you can check out my blog post on "The Bot Scare" which is not peer-reviewed but I try to cite lots of sources and make a decent argument that most of this kind of research is pretty flimsy.

Also worth looking at is this informal audit of a few "bots" that were identified by these researchers back in April, some of which are humans with faces and lives who post videos of themselves like, talking and living and stuff

COVID-19 is all anyone is talking about, and unless we posit that there are more bots than people out there on social media, there needs to be extremely good data to make a claim that half of all conversation about COVID-19 is from bots. The burden of proof is huge and not met.

But even on top of that, use your brain: of the human people in your life, your friends and family and coworkers who you follow on Twitter... what percentage of their posts have been about the pandemic in one way or another for the last 2 months? For me it's easily above 50%.

The short of it is: knowing what we know about the study, which is very little, it seems like these researchers have in the past used a very loose and nearly useless definition of "bot"

NPR is promoting this article again. Without access to the study we have no way of knowing how "bot" was estimated or measured, we simply have to go on the reputation and past research of this lab, which I dug into last night here: https://twitter.com/tinysubversions/status/1263675864568356864?s=19

RT @thehowie: Overnight, https://covidexitstrategy.org updated their map to reflect #Mississippi deterioration -Southeast now a solid block of re…Quote Tweet(((Howard Forman)))@thehowie · 14 Jun#Mississippi is a small (pop) state with poor testing, increasing positive rate, HIGHEST hospitalizations per capita (and ), and under-resourced. Another state to watch closely. @LeeZurik

Feels like England has lost the plot & can't see the wood for the trees. Obsessing over 1m v. 2m (tbh, restaurants/pubs not financially viable at either), mandatory masks -> instead of larger questions of what is the objective & necessary strategy? How best to prepare for winter?

Countries where citizens report higher openness to diversity are more likely to become democratic. The authors say that this trait is important because 'it predicts peaceful coexistence of competing viewpoints'.

In the wake of recent events, I keep thinking about this paper, published by @damianjruck et al. earlier this year. https://nature.com/articles/s41562-019-0769-1… short thread:

This paper challenges the widespread assumption that we can create democracies by introducing democratic institutions, and that we can inculcate support for democracy with the right set of societal rules. Spoiler: we can't.

There is no enshrined rule or law that democratic nations must remain democratic. Democracy is a political choice - it can be swept away with the tide of public opinion. These findings make me worry about the fate of some countries - including my own - over the coming years.