Adventures in creative graphing


Every year, I have to teach students about how to properly chart data. I’ll have to use this as an example of how to do it badly.

(I know, this isn’t literally put out by the White House, it’s mocking the Republican downplaying of COVID-19 deaths.)

Comments

  1. says

    If it wasn’t for the Benghazi being higher than the Ebola that could almost be an accurate log plot with the X axis at the top and the colours being what isin’t……Aw forget it, silly joke.

  2. davidc1 says

    Over here the govt is hiding the real death total’s by not including people who die of covid 19 in care homes or their own homes .
    But never matter ,bojo is well and having a nice holiday at his pad in the countryside .

  3. says

    It is also out of date already, since today over 22 thousand Americans died of COVID-19. And of all the disasters listed, the COVID-19 is the one that could be best avoided/mitigated, there was plenty of time to prepare and enough information about what happens if decisive action is delayed too much.

  4. Snarki, child of Loki says

    Oh, your face is going to be really red, when Trump retweets your graph to his adoring MAGAts, then spends a good 15 minutes of his daily Pravda pandemic reality TV segment using that graph to polish his ego.

  5. komarov says

    This is all well and proper, they just forgot to display the political weighting factor, colloquially known as the Obama constant (ironic, since it keeps changing). I’ll leave it as an exercise for the reader to work that one out from the data.

  6. blf says

    Also note the four thug disasters were all (broadly speaking) USAian (i.e., domestic), whilst the two terrible Obama-era disasters were (again, broadly speaking) in nasty furrin shiteholes full of brown & black people but none of hair furor’s properties.

  7. raven says

    Over here the govt is hiding the real death total’s by not including people who die of covid 19 in care homes or their own homes .

    AFAIK, the USA isn’t hiding assisted living home deaths (roughly 14% of the total deaths) or deaths that occur at home.

    These can though be quite high and are often not counted for a lot of reasons.
    In Italy during the height of their pandemic around Milan, the at home deaths were 2 to 4 times the hospital reported deaths.

    We are seeing a huge number of at home deaths here in the USA. Some days it is 3/4’s or so of the total Covid-19 deaths that day.
    I’m not at all sure why that is.
    People are getting sick and dying without even making it to the ER.

    Some of these at home deaths get counted in the Covid-19 totals and some don’t.
    It is though, quite clear that the official death totals are under counts by a long ways.

  8. whheydt says

    Re: raven @ #7…
    Part of the reason, as Graydon Saunders put it in one of his Commonweal books…you don’t triage the dead. In this particular case, you don’t test the dead, especially when those who would do the testing are already too busy dealing with the (still) living and test kits may be in short supply.

    But, yes, like virtually any major disaster, there will be a significant undercount of victims of the disaster, regardless of any contributory factors. That said, sometimes the undercount is–politically–deliberate. Read up on the death toll from the 1906 San Francisco Earthquake. Officially, at the time, about 1000. In fact, as dug out much more recently, in excess of 3000.

  9. says

    @Raven:

    AFAIK, the USA isn’t hiding assisted living home deaths (roughly 14% of the total deaths) or deaths that occur at home.

    What would you consider “hiding” deaths that occur at home? Since we don’t test the bodies of persons who died at home, and we don’t add any of those deaths to the COVID-19 total (even though statistically some of them have to be, and in places like NYC there’s good enough historical data on average at-home deaths + a large enough number of deaths to make good estimates)?

    If we released our data and reported our date in the media with appropriate qualifications (like with an asterisk and an end note “deaths at home have not yet been quantified, thus this number should be taken as a minimum with the total being as much as X% higher”), then I’d say we weren’t hiding them. As it is, “hiding” certainly feels like an accurate (if colloquial) description to me.

    So I’m just curious as to why you feel the way you do and what you would consider to be “hiding” deaths.

  10. raven says

    So I’m just curious as to why you feel the way you do and what you would consider to be “hiding” deaths.

    The deaths are reported by the 50 states, not the federal government.
    Some of them might try to minimize the death count but not all 50 of them.

    And some states do test the at home dead.
    That is why we know that many Covid-19 deaths occur at home.

    If you want to invent some huge conspiracy theory over this, feel free.
    I’ve got a million more important things to think about and a busy day ahead of me.

  11. blf says

    Here in France, there is a known — and acknowledged — glitch that deaths in nursing homes due to Covid-19 weren’t being counted as Covid-19 victims. They™ are working on fixing this, albeit to-date the reporting of nursing home deaths is erratic as various problems are sorted out. (As noted below, the current numbers are now roughly consistent with other, more robust, figures from other countries, suggesting the problems are now mostly ironed out.)

    For more on the situation in the UK w.r.t. nursing home deaths, see this excerpt in poopyhead’s current [Pandemic and] Political Madness All the Time thread; unlike France, it is unclear if attempts are being made to fix the UK’s “glitch”.

    Some States, like New Mexico, are apparently dealing with the nursing home situation correctly.

    Finally, Half of coronavirus deaths happen in care homes, data from EU suggests:

    Figures from Italy, Spain, France, Ireland and Belgium suggest UK may be underestimating care sector deaths

    About half of all Covid-19 deaths appear to be happening in care homes in some European countries, according to early figures gathered by UK-based academics who are warning that the same effort must be put into fighting the virus in care homes as in the [UK’s] NHS.

    Snapshot data from varying official sources shows that in Italy, Spain, France, Ireland and Belgium between 42% and 57% of deaths from the virus have been happening in homes, according to the report by academics based at the London School of Economics (LSE).

    Published official data for care homes in England and Wales are believed to significantly underestimate deaths in the sector, with the Office for National Statistics only recording 20 coronavirus-related deaths in all care homes in the week ending 27 March. New figures are due out on Tuesday, but are unlikely to be up to date.

    […]

    In its first review of early international evidence, researchers from the International Long Term Care Policy Network (LTCPN), found the most robust data was available in Ireland where, as of Saturday, 54% of deaths from coronavirus occurred in care homes, according to centralised government figures.

    In Italy, a government survey covering deaths in 10% of the nations’ care homes suggested 45% of all deaths in that country could be in care homes, while central government data in France released at the weekend showed the same proportion of the 13,832 deaths in that country happened in care. In Belgium, health ministry figures showed 42% of Covid-19 deaths happened in care homes and in Spain, leaked regional government data suggested 57% of the country’s death toll from 8 March to 8 April was from care homes.

    […]

    There have been three epidemiological studies of Covid-19 in care homes carried out by authorities in Seattle, one of which found that by 18 March, of 101 care home residents who contracted the virus in the preceding weeks, 34 died.

    […]

    Dr Michael Head, a senior research fellow in global health at the University of Southampton, […] cautioned that it remains hard to draw conclusions on precise numbers of deaths of elderly people in nursing and residential care around Europe and said that given there are few tests available, very few suspects cases will end up as confirmed cases. […]

  12. raven says

    “I’m not at all sure why that is.”

    I’ll add here that what I’m not sure of is exactly why so many Covid-19 deaths are occurring at home.
    Everyone does have the option of heading to the nearest ER, calling 911, etc..
    We do know that getting to the hospital makes a huge difference in survival rates as well.
    They can’t save everyone but they can save a lot of people.

    We do know that people with Covid-19 can take a turn for the worse and develop serious pneumonia quite rapidly.
    It could be that people are just getting sick and fading before they can decide to go for medical help.
    Or they have underlying conditions that hinder their ability to make such decisions, i.e. dementia, various other disabling conditions.

  13. blf says

    @12, The insurance-led workers-only dysfunctional “health care system” lurks… who is going to pay once the patient is no longer considered an emergency case? That is probably preying on many people’s minds.

  14. says

    @raven

    If I get this I am not going to hospital. I don’t have the 73,000+ dollars to pay for treatment. There’s part of your answer.

    New York state/city is not testing at home deaths at this point and they haven’t yet reported the figures of suspected cases. They announced on the 8th that they would but haven’t gotten to it yet. Given that NYC is driving the national outbreak (about 50% for cases and deaths) that is a huge oversight in the numbers.

    I don’t think “hidden” is the right word either for what is happening here but oversight, under reported, incompetence are all germane.

  15. blf says

    @14, Thanks, that’s a somewhat more concrete instance of my speculation in @13. (Ok, I’ll admit I like to see some citations, to pin down the point.)

  16. says

    Correction: the Tristate area-New York, New Jersey, Connecticut- makes up about 50 of cases and deaths. New York State is only about 1/3 but the New York City metro area is the epicenter at the moment and that’s what I was thinking of.

  17. unclefrogy says

    my guess as to why there is any suspicion about the reporting of deaths and the inaccuracy is connected the history of reporting negative information around disasters and war which still goes on today. They are often touted for political advantage . it is an example of overly legalistic thinking applied in the narrowest way possible.
    uncle frogy

  18. wzrd1 says

    New York is “hiding” COVID-19 deaths, as there are not enough tests to test the dead for a definitive diagnosis, so the death certificate won’t list Coronavirus as the cause of death. Keeps Trump’s numbers down.

    In other news, I heard on grate authority on Twitter (yeah, I know) that hospital parking lots are only half full, so the hospital only has half of the beds filled.
    Because, we all strap our cars, whether or not we own one, on top of the fucking ambulance when being transported for hypoxia and dyspnea to the emergency department, amIright?!
    Just further evidence that folks, it’s a bad idea to poke baby in the fontanelle repeatedly, it causes brain damage.

    As for bad graphs, I once did one for an information security seminar for a business case for a risk mitigation for a recalcitrant management in regards to any changes. The mitigation was an extremely low cost mitigation, so I made the graph with that in mind and the risk acceptance chart still used that same low dollar metric, with a cost of one billion dollars if the event to be protected against did occur (based upon the US CENTCOM cyberattack cost of remediation and the scope of the network mitigation was well, the same network). The chart had a fanfolded extension for the billion dollar cost.
    Despite the ludicrous fanfolded extension going up to the ceiling and back down to a fanfolded pile on the floor, the exercise, once the laughter died down, was deemed a resounding success.
    At times, the absurd gets the point across far better than scaling things down to sane and more easily ignored or overlooked levels.
    It’s also a barely acceptable means of being a dick, when dealing with dunderheads.

    The current horror today is, as of 16:45 EDT, is 1.9 million ill, 118452 dead globally. 582587 sick, 23078 dead, 2.8 million tested, 93631 hospitalized in the US. We’ve beat every nation now in cases and deaths, chalk one more up for American Exceptionalism!
    Sleep, used to sleep well. Now, it’s as if I felt a great disturbance in the Force, as if millions of voices suddenly cried out in terror and were suddenly silenced.

    If one wishes to really get irritated, read the NYT’s excerpts from the “Red Dawn Breaking” e-mail chain. I’m still spitting fire and I’m only 1/3 of the way through.

  19. blf says

    me@15, Here’s a citation which broadly backs up Mike Smith@14’s comment, “If I get this I am not going to hospital. I don’t have the 73,000+ dollars to pay for treatment”, If you’re hospitalized with COVID-19 and don’t have insurance, you’ll likely owe $73,000. One chart shows how broken the system is for the most vulnerable (27-March-2020):

    ● Uninsured Americans with COVID-19 would pay an estimated average of $73,300 for a 6-day hospital stay, according to a recent report by FAIR Health.
    ● There are 27 million Americans without insurance.
    ● COVID-19 symptoms range widely and treatment could vary, so costs could realistically be far less — or more.
    […]

    A recent report by healthcare cost database FAIR Health, which looked at the economic impact of the coronavirus pandemic on the healthcare system, found that a six-day inpatient stay for treatment of coronavirus symptoms would cost 100% of the price for someone without insurance: a whopping average of $73,300. (The same would apply if you stay in a hospital outside of your insurance network and your insurance plan doesn’t have out-of-network benefits.) There are 27 million Americans without insurance, despite the passage of the Affordable Care Act.

    The analysis considered nearly 18 million people under age 65 who used employer health plan claims.

    [… the graphic mentioned in the title…]

    Those with insurance are estimated to have an average medical bill for the same duration stay of $38,221, but that doesn’t mean that’s how much they’ll pay. That is the rate negotiated between your insurance plan and your in-network health providers. How much you’d actually pay depends on your insurance plan.

    [… E]ven if you aren’t hospitalized for COVID-19, testing and treatment could still cost a pretty penny. Danny Askinsi, based in the Boston area, told Time magazine that she tested positive for COVID-19 and is uninsured. Her testing and treatment involved three trips to the ER before she was tested, resulting in a nearly $40,000 medical bill.

    And two people who didn’t test positive, but received other related treatment and testing, received bills for over $3,000. Add up all these high medical costs for matters of life and death during a pandemic and you see a broken healthcare system at a time when Americans literally can worst afford it.

    That broadly seems to back-up my speculation in @13 that one possible answer to raven@12’s puzzlement “I’ll add here that what I’m not sure of is exactly why so many Covid-19 deaths are occurring at home” is “The insurance-led workers-only dysfunctional ‘health care system’ […]”.

    In contrast, here in France, President Emmanuel Macron announced earlier today that by May 11th, everyone who is displaying symptoms will be tested. He didn’t mention costs, but with the French health system, it will either be for free or (for those who can afford it) a very small cost (a few euros); similarly for any needed hospitialisation. (Yes, France has also gotten off to a slow start on testing — for which Macron apologised (can one even imagine impeached quack hair furor doing that?) — but that is clearly being addressed.)

  20. vucodlak says

    @ raven, #12

    “I’m not at all sure why that is.”

    I’ll add here that what I’m not sure of is exactly why so many Covid-19 deaths are occurring at home.
    Everyone does have the option of heading to the nearest ER, calling 911, etc..

    Several reasons.

    The big one is that no one wants the hassles and enormous bills that come from dealing with our for-profit medical system, so they put off seeking care as long as possible. Often, they wait too late, and die before they ever see the inside of a hospital. Always, always remember that this is by design- health insurance companies want people to be reluctant to seek costly medical attention.

    If the Trump administrations mishandling of this crisis is the number one cause of the catastrophic toll this virus is taking on the US population, the health insurance industry is a very close second. To make it even more disgusting, the industry is likely to see a major windfall from this plague.

    The next one is that hospitals critical care facilities are overwhelmed. It’s not that the hospitals are all packed to the rafters, it’s that the relevant resources for treating covid-19 patients are already at or over capacity. Hospitals are laying off employees left and right because so many elective and non-urgent procedures have been canceled, and those employees aren’t able to assist with the pandemic for whatever reason.

    If you go to the hospital with covid-19, but don’t require urgent care, you’re likely to be sent home to self-quarantine. So if you do end up requiring critical care, you’re likely to hesitate too long (because they’ve already sent you home once, and you don’t want a huge bill for something that your insurance might deem “unnecessary”) and die before you get back.

    Basically, it’s a combination of incompetence on the part of our government, unpreparedness on the part of our health care system, and sheer, murderous greed on the part of our health insurance industry. Not to belabor the point, but that third thing is largely responsible for the second thing. Preparedness costs money.

  21. chrislawson says

    $73K!!! Here in Australia if you go to a public hospital with COVID19 and need ventilation and ECMO and are lucky enough to recover, you will receive a bill for zero dollars. At most you will have to pay $40-50 for your first prescription medications at discharge. And yet half of Americans seem to think this is impossible to achieve in a health care system.

  22. chrislawson says

    jkrideau–

    Thanks for that interesting link. I fail to see how Fox’s graph can be anything other than deliberate finagling to disguise the severity of the growth rate. It makes no sense at all as an error of incompetence.

  23. jrkrideau says

    @ 25 chrislawson
    I fail to see how Fox’s graph can be anything other than deliberate finagling
    I think it in likely but I do not think basic incompetence by some innumerate grapic artist with a good drawing program plus a couple of helpful suggestions from an equally innumerate boss.