Across the lines

I haven’t really come across anything to write about recently. I’ve been thinking a bit about models and how they are used to inform decision making. I’ve been thinking a bit about the use of scientific advice. I also had an interesting discussion on Twitter with Jean Goodwin about what to expect from scientists who engage publicly. I may write a bit more about this at some stage, but my concern with this is that we should – in my view – be careful of constructing a narrative that then allows people to blame scientists if the decisions we end up making aren’t regarded as optimal (I will admit that the current situation has confused me somewhat).

For some reason, I’ve been listening to quite a lot of Tracy Chapman recently. I remember listening to her music a lot many years ago, and only came across it again recently. The song below seems quite apt, unfortunately.

This entry was posted in advocacy, Personal, Philosophy for Bloggers, Policy, Politics and tagged , . Bookmark the permalink.

20 Responses to Across the lines

  1. Not-in-my-name says:

    You refer to “we”, when I assume you mean the government. It is the government that is using scientists to advise and if they make decisions that we do not like we are entitled to blame both the government and their scientific advisers. If I personally use the advise of a scientist and it is wrong then I am entitled to blame the scientist. Are you saying people cannot blame Andrew Wakefield for following his advice not to have the MMR vaccine?

    However, it depends on the nature of advice. The great Thomas Sowell says there are no solutions, just trade offs. This is what we have seen with the advice over the coronavirus. The government made decisions that they believed would save lives due to the virus, but they ignored completely the consequences of their advice that will result in deaths for other reasons. We know that the NHS has stopped all other treatment and this will result in deaths in the longer term. They will now have a a horrendous backlog of patients needing treatment with the result of more deaths, and this is without considering the mental health issues resulting from the lockdown. Humans need freedom, not government imposed limitations on freedom.

  2. Not-in-my-name,
    By “we” I mean society, which often does then mean that the decisions are made by elected representatives (or, governments).

    I think it’s a complex issue. Certainly, if a scientist, or group of scientists, give advice that is wrong and that they should have known was wrong, then of course they can (and should be) criticised. The complexity comes in when you include all the other factors that will influence decision making. It’s quite rare that the scientific evidence defines what decision should be made (sometimes it might be obvious, though). There are many other factors that play a role (economic impact, impact on our health, etc). As you yourself indicate, our current response to the coronavirus threat has other consequences and we do need to try and balance all of these when making decisions. The responsibility for taking all of this into account should really lie with our decision makers, not with one group of advisors. Having said that, some of the recent rhetoric has made me wonder if those who give advice recognise this subtlety.

  3. JCH says:

    Who has ended up being free? The people who got locked up. Who is trapped in virus hell? The people whose leaders were lazy and said “but we’re free.”

  4. Joshua says:

    name –

    > You refer to “we”, when I assume you mean the government.

    Your logic is flawed if you’re referring to a country where the government is elected. Certainly, in most governments the election process is flawed; for example, in the US Republicans and whites and men and people who live in less populous states are over represented. But even with that, the government is still an approximation of “we.”

    I gather you would like to have even more disproportionate power in government decision-making – beyond the disproportionate power I presume you already have by race, ethnicity, ecomomic status, and gender. Of course, one way for you to have that would be to form your own country. Or more easily move somewhere where you’d be even better able to leverage your priveledges to access greater influence and power. I can suggest a few counties if you’d like.

  5. I listened to Sir Mark Walport a few years ago in a conversation about the role of Chief Scientific Adviser (a post he has held), which was very interesting


    [This episode was recorded on July 21, 2016 in front of a live audience at Caspary Auditorium at The Rockefeller University.]

    He said that any policy must look at a problem through 3 different lenses:
    – Evidence lens
    – Deliverability lens
    – Values lens

    and that science can only help with the first of these.

    He made the point that trust in science is very context specific: Science can say anything about the Higgs Boson and be believed, but on an issue like embryology, values kick in and there will be much less trust.

    He also makes a strong distinction between ‘rollable’ questions and non-rollable questions.
    “does extra CO₂ in the atmosphere lead to increased global warming?” is a non-pollable questions (answer: yes); whereas “should UK focus on renewables or nuclear power to decarbonise the grid?” is a pollable question (answer: Brits much prefer renewables, by a wide margin).

    Scientists need a special range of skills to be able to do the advice job, above and beyond their scientific skills. John Ziman explored the differences between scientific discourse and political debate in his paper (2000) “Are debatable scienti􏰜fic questions debatable?”

    Click to access Ziman.pdf

    He explains how complex most scientific questions are, with rarely a simple resolution, and conducted in a way quite different to political debate (yet no less argumentative!). The two styles sit awkwardly together.

    Yet public and political discourse (especially on social media, but in newsprint, and parliament too) often expects a binary answer: yes or no, right or wrong. Shades of grey are often not tolerated, and if you don’t ‘choose a side’, expect to get caught in the crossfire.

    I haven’t read the belatedly released SAGE Minutes yet but I expect there will have been lots of discussions on points where Walport’s lenses (Evidence, Deliverability, Values) sit uncomfortably alongside each other.

    at some point, I imagine a fly on the wall, hearing …
    “we need to do test, trace and isolate as soon as possible”
    “agreed, but we need to recognise the constraint that the test capacity is limited at the moment, so we’ll have to wait till we have flattened the curve enough, to reduce the testing demand, but also build up capacity; meanwhile we cannot avoid a lockdown”
    “can someone answer this – how well will the public will comply and how does change the numbers?”
    “we ran some sensitivity analysis, and we need very high compliance to make it work”

    Leading to a messy compromise set of ‘options’ and scientists NOT the ones with the authority to choose which ones.

    The scientists didn’t choose a context where Governments had failed to take on board prior recommendations over some years, to build capacity in PPE, etc. So the advice is very context dependent.

    It is highly disingenuous of politicians to say they are ‘following the science’ when that is just one element in the decision making, and where a poor starting position (e.g. the lack of prior investment in pandemic responsiveness) is neither something they influenced, nor can change.

  6. Errata: “pollable” not “rollable” (I fixed it but spellchecker seems to have the ability to have multiple goes at screwing up text grrr).

  7. Richard,

    He said that any policy must look at a problem through 3 different lenses:
    – Evidence lens
    – Deliverability lens
    – Values lens

    and that science can only help with the first of these.

    Yes, this seems pretty obvious (although, I’d argue that research can help with the second too). I am, however, constantly amazed that we seem to regularly have discussion that suggest that either this isn’t obvious, or that scientists don’t realise this.

  8. russellseitz says:

    In 2005 Ziman observed that
    “scientists are no longer lonely seekers after truth, but have emerged with multiple obligations as technical and military experts, entrepreneurs, managers, political advisers, publicists, and educators, as well as ordinary citizens…
    The personal preferences of scientists are now transformed and often under the control of mammoth institutions – great universities, a tangle of granting agencies, huge defense establishments, and global corporations. Rarely do scientists work alone in isolated laboratories. They are linked together in intricate networks,”

    But all those institutions and entities now employ networks of publicists and PR men who monitor social media with far greater zeal than they read science. The use of focus groups to pitch science to the public in Ziman’s time was disturbing enough- what would he make of today’s deployment of social media algorithms and the public engagement of broad editorial collectives like Covering Climate Now?

    The flood of real time covid research publications is taking its toll on pundits as well as scientists trying to keep their heads above water:

  9. Dave_Geologist says:

    Indeed russell. Outliers excepted, and some like Santa Clara can be excepted on their own flawed methods not just for being outliers, the IFR stays stubbornly stuck at around 1% and the asymptomatic rate at about 50%. Just like the early studies based on that supposedly fake Chinese data and supposedly “special” Diamond Princess data said, months ago.

    It’s almost as if epidemiologists know what they’re doing. Just like the climatologists who said decades ago that ECS was about 3C knew what they were doing.

    We could have fooled ourselves three months ago that it was possible to pursue herd immunity and shield the vulnerable. Not any more, with deaths in care homes stubbornly representing about half of all deaths, even in the country which pursued herd immunity and knew it had to protect care homes.

    As with AGW the basics are simple and the rest is detail. Kill 0.5-1% of the population on the way to herd immunity. Or pursue a Plan B.

  10. an_older_code says:

    and of course the counter factual is that if there were no outliers (either way) and every study “agreed”

    then we would get the “its a conspiracy”

    although tbf that does not stop the conspiracy crowd anyway

  11. Dave_Geologist says:

    N-i-m-n, the relevant deaths to compare with unintended-consequences deaths are those which would have occurred without a lockdown, not those which occurred despite the lockdown. About half a million in the UK, now we know that huge numbers of asymptomatic cases and huge population-infection rates and herd immunity are fairy stories. We couldn’t be sure of that at the time the call was made, but with hindsight we now know that the right call was made, and indeed that it should have been made sooner.

    Because there is so much variation in the way deaths are counted, or rather allocated to Covid-19 vs. other causes, it’s best to look at excess deaths normalised against a seasonally-adjusted five-year average. For example, the Euromomo figures. These are compiled from all registered deaths, with Euromomo doing the calculation on data submitted by partner agencies like Public Health England. Country graphs are at the bottom and you can use the slider bar to zoom in to 2020. Look at the huge humps for England and Spain: It will take a humongous increase in collateral deaths due to lockdown to compensate for that. I’d take the other side of that bet. And remember, these are with-mitigation Covid deaths, not unmitigated deaths. In fact, a number of countries have recently dropped below the long-term average. You might argue that that shows there are people who’d have died anyway of flu or whatever a few months later. But again, remember, this is with Covid mitigation and we’re already late in the flu season. Look at previous years to see when the flu peak occurs. Generally about four months earlier. And indeed I read recently that in the UK non-Covid flu, pneumonia and other winter deaths like slips and falls have been lower than usual since lockdown. IOW far from putting them at risk, shielding vulnerable people has protected them from normal winter-excess-death risks. Of course we may see a massive tail of excess deaths before the second wave and then I can eat my hat. But actually a lot of those should already be happening: untreated heart attacks, urgent but non-emergency cancers. And someone who needs an operation some time in the next few years is safer at home while Covid is in the hospitals.

    Comparing to flu, we can see that all the badly-hit countries are hit much worse than in the 2017 or 2018 flu seasons. And that’s with heavy Covid mitigations vs. just vaccinating the vulnerable (I think a few countries had school closures at the peak of a flu outbreak, but that’s all). The bad flu year depends on where you are: 2016-17 was cold in central Europe, 2017-18 was cold in NW Europe. Some countries got both (England, Spain), some one or the other (Scotland, Greece). The Scandinavian countries are immune to cold snaps: as my Norwegian relatives say, there’s no such thing as bad weather, just the wrong clothing. The countries which were worse hit by past flu seasons than by Covid all took strong lockdown measures. Those who say it’s no worse than flu are right: but only if you control it.

  12. Steven Mosher says:

    ” asymptomatic rate at about 50%. ”

    Diamond Princess

    3,711 people tested
    634 tested positive


    Vo’Euganeo, (50Km west of Venice)

    All 3,000 people in village tested

    50 – 75% asymptomatic

    Asymptomatics infectious
    represented ‘a formidable source’ of contagion

    328 adults in Shanghai


    328 adults were diagnosed with COVID-19 in Shanghai

    Among them, 13 patients were asymptomatic at time of diagnosis

    No subjective symptoms 2 weeks before admission

    All of the 13 (4%) them were close contacts of the confirmed cases

    Japanese nationals evacuated from Wuhan

    565 citizens had been evacuated
    30.8% (7.7– 53.8%).
    30 days after evacuation

    Asymptomatic and Pre-symptomatic SARS-CoV-2 Infections in Residents of a Long-Term Care Skilled Nursing Facility – King County, Washington
    CDC 76 residents

    categorized as asymptomatic or symptomatic at the time of testing and for previous 14 days

    23 (30%) residents, antigen positive and symptomatic
    13 (57%) were asymptomatic

    10 of these 13 previously asymptomatic residents developed symptoms within 7 days

    Therefore 3 (4%) out of 76 truly asymptomatic

    RT-PCR) values indicated large quantities of viral RNA in asymptomatic, pre-symptomatic, and symptomatic residents

    Symptom based screening on a particular day could fail to detect half of positive cases

    Population screening, 57% symptomatic, 43% asymptomatic

    CDC, Interim Clinical Guidance for Management of Patients with Confirmed Coronavirus Disease (COVID-19)

    13% of positive children asymptomatic

    Risk of transmission is thought to be greatest when patients are symptomatic since viral shedding is greatest at the time of symptom onset and declines over the course of several days to weeks.

    81% of infections show mild symptoms
    14% exhibit severe symptoms
    5% fall critically ill

    Asymptomatic coronavirus infection: MERS-CoV and SARS-CoV-2 (COVID-19)

    The extent of asymptomatic MERS-CoV (2012) cases is about 9.8%

    How much is covid-19 spreading via asymptomatic versus symptomatic infections? (4th May)

  13. Dave_Geologist says:

    Asymptomatic in the sense of none or mild Steven, the sort of thing you wouldn’t think about in a normal year. Mildly feverish but only had one of those things you press to your forehead and it didn’t go red. Bit of a cough but no different to previous coughs I’ve had. Generally feeling under the weather. Upset stomach/diarrhoea/vomiting (Mary Wakefield). Etc. Part of the variation is probably real (age/health of population), part self-selection/non-blinded. Tested positive as a contact or in a survey. Did you have symptoms? Well now that you mention it… Would you have self-isolated based solely on those symptoms? No.

    That’s one of the things the group behind the UCL symptom-tracking app are banging on about. They have a list of ten or twenty symptoms, most of which are not on the official list of three or four. They think that with better guidance more of those “asymptomatic” cases would be identified and silent spreaders could be take out of circulation. The problem with somewhere like the UK is that with our current infection load and testing capacity, broadening the list would result in huge numbers of people who’d test negative self-isolating.

    In countries which have the capacity to quickly chase down every case, there could be merit in having a catch-all symptoms list. What does Korea do?

    Still, I was thinking mainly of whether the rate was credibly at a level where we could say we’re almost at herd immunity and just need one more push then we can all relax. For that purpose it doesn’t matter whether the asymptomatic rate is 15%, 50% or 75%, or pretty much anything less than 95%.

  14. Ben McMillan says:

    How do scientists avoid ‘constructing a narrative that then allows people to blame scientists if the decisions we end up making aren’t regarded as optimal’?

    Clearly anyone involved at any state of the decision-making process is in the line of fire, so best you can hope for is less blame.

    There is also the issue that giving correct (or best-judgement) advice might be more important than avoiding future blame. I know doctors spend a fair bit of time worrying about whether their advice is really the best for the patient, or the one that prevents them getting blamed.

  15. You could right about this:

    Antarctic Ocean Reveals New Signs of Rapid Melt of Ancient Ice, Clues About Future Sea Level Rise

    quote from Eric Rignot: “The most important message to take home is that the current projections are too conservative. We know it,” he said. “The real drama in all of this is that the faster rates of retreat may turn out to be the most probable in some places, and as of now we do not know where and when.”

    Some of us have been saying the projections are too conservative for years. That really has been quite obvious, but the “cooler heads” always like to chide the alarmists/realists who say, hey, the projections are too conservative. Now scientist Eric Rignot says that is the most important message to take home: the current projections are too conservative. We know it.

    I wish I could take some satisfaction about being right on this matter, but it’s a pyrrhic victory.

    What about it, ATTP? Are the current projections too conservative? Do you know that?

  16. russellseitz says:

    Dave the Geologist says

    “It’s almost as if epidemiologists know what they’re doing. Just like the climatologists who said decades ago that ECS was about 3C knew what they were doing.”

    Dave , since epidemiologists tend to be rather more disinterested than wildcatters, that is even less surpising than the ECS outliers being authored by folks who either died a century ago, like Greta Thunberg’s great-great uncle Arrhenius, or solid scientists who give well paid pep talks to coal companies, like Dick Lindzen.

    Taking the 57 flavors of ECS estimates as Bayesian priors, it looks like the climatologists are mon an heuristic roll.

  17. Everett F Sargent says:

    World = World – CN, RoW = World – (CN + EU + US) and SA+MX = South America + Mexico
    7-day rolling mean for dailies, 1-day for cumulative

    EU, US and SA+MX have ~flatlined over the past week for daily deaths. RoW growth seen in PK+IN+BD+RU (not shown).

    2nd near term peak very probable and much sooner than I would have expected just a week ago.

    In my homeland, with a pandemic, 40+ million unemployed and 10 days after a killing of a black man, will Small Hands follow in the footsteps of Nixon?

  18. Mal Adapted says:


    If I personally use the advise of a scientist and it is wrong then I am entitled to blame the scientist. Are you saying people cannot blame Andrew Wakefield for following his advice not to have the MMR vaccine?

    You may be confusing “the scientist” with other people, who have transparently non-scientific agendas. Wakefield was a medical doctor who published fraudulent research for personal aggrandizement. When his fraud was documented by a freelance investigative journalist, he was formally censured and cast out by the medical profession, as represented by its licensing board and flagship peer-reviewed journals. Although the field of biomedicine may offer financial incentives to bias research results, the norms of scientific culture firmly reject outright fraud.

    OTOH, anyone who takes a high-school science course recognizes that while science is primarily concerned with reducing errors of observation and analysis, some uncertainty is ineradicable, so that scientific conclusions are always tentative and conditional. Undue confidence is far more likely to come from politicians and other self-interested non-scientists seeking to exploit science. So before blaming “the scientist”, make sure they actually said what you think they said, and not what somebody else claims they did. And if an ostensible “scientist” does make a consequential claim out of their own mouth, you still have to be skeptical. Read the primary peer-reviewed source yourself. Find out what other specialists in the discipline are saying about it. Then you, and only you, must ultimately decide how much to rely on it. Because despite rigorous empirical training and the incessant, unsparing scrutiny of their peers, scientists are people. They make mistakes in their work, which their peers discover sooner or later. It’s the global culture of science that makes the accumulating body of justified knowledge self-correcting.

    BTW: setting aside actual fraud, what makes you entitled to blame scientists? Do you sign their paychecks? What do you think their contract with you is? When did they promise that their advice was 100% correct and accurate? It’s up to their professional peers to judge how reliable the advice is, not you in any case. Not to put too fine a point on it, but your sense of entitlement is suspect.

  19. angech says:

    Perhaps the recent Lanclet and the American studies on coved are worth a comment, ATTP.
    Highly respected journals caught out or shoddy peer review?

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.