Talk:COVID-19 pandemic in the United States/Archive 14

Latest comment: 4 years ago by 2605:E000:100D:C571:E0A0:848E:577F:F036 in topic Graphs in Number of U.S. cases by date section
Archive 10Archive 12Archive 13Archive 14Archive 15Archive 16Archive 17

Variations in case totals by countries

Is Europe ringing a false alarm? Headline: AP, Aug. 8, 2020.

U.S. tops 5 million confirmed virus cases, to Europe’s alarm
With confirmed coronavirus cases in the U.S. hitting 5 million Sunday, by far the highest of any country, the failure of the most powerful nation in the world to contain the virus has been met with astonishment and alarm in Europe. (L.A. Times)

Maybe it is a false alarm. According to the WHO, about 80% of Covid infections are asymptomatic. Some other U.S. estimates feel that it could be closer to 50%. In any case, according to Harvard experts, "All of the best evidence suggests that people without symptoms can and do readily spread the virus" and "evidence suggests that people may be most infectious in the days before they become symptomatic."STAT.

However, the U.S. seems to be the only country in the world that is testing those with or without symptoms, as stated in the CDC site. While according to the European Centre for Disease Prevention and Control (ECDC), which recommends testing procedures for the continent, they only recommend tests for people after they show symptoms.[1] And in less developed countries, tests are less available for those even with symptoms.

So based on those facts, if the U.S. only tested those with symptoms, as they do in Europe, it might not have found from 50% to 80% of infected persons, giving it 50% to 80% fewer reported cases. But because the U.S. does test asymptomatic people, who are then more isolated if infected, they are less able to spread the infection. While in Europe, because asymptomatic and thereby untested people are at least as likely to spread the virus, it seems an American visiting Europe is more at risk than the other way around. So is the headline a false alarm based on wrong facts? --Light show (talk) 22:55, 9 August 2020 (UTC)

U.S. has reached one of the highest daily test rates per thousand pop in the world, per Our World In Data (OWID)[2]. (Note the test data is not available for China; are they not testing?) At the same time, many U.S. states are operating more like Sweden now that they have reopened, so there is no surprise the cases went up, but the important thing is that this does not translate into these U.S. states having the kind of very high mortality rates that we saw in New York City and New Jersey, as per COVID-19 pandemic in the United States#Mortality rates. On yet another note, instead of taking the meaningless absolute figures, we can look at per capita rates in Our World in Data[3] and note that U.S. has indeed one of the highest confirmed case rates in the world, 1.5% of the population positive, whereas e.g. Spain has 0.67% of the population positive; that again is explained by both the high test rates (U.S. has reached triple the daily test rate of Spain) and by the reopening in many states. (Let us keep in mind that the actual number of people who have had the covid infection and now have immunity either via antibodies or via T cells is much higher than what the confirmed cases show.)
Let us also note the phrase "Europe’s alarm": that implies that either Europe is a person or that there is some kind of aggregate alarm across European Union, but that has not been substantiated by the article. What the article does is that it quotes some people in Europe who are critical of the U.S. response and would like to see hard lockdowns, thorough trace-and-isolate and mandatory wearing of face masks; but surely there are enough people in Europe that can be quoted who praise the reopening in the southern U.S. states while criticizing the response in New York City, for which sources report that covid-positive patients were sent from hospitals to nursing homes.
Furthermore, European country rules that regulate entry of foreign nationals would not be concerned with total cumulative numbers from the beginning of the pandemic but rather with current active numbers or the rate of recent new cases (daily, last 7 days or last 14 days) per capita; the article chose to report the cumulative total instead. On the other hand, new daily cases per capita are still high in the U.S. per OWID[4], which may lead some countries to regulate U.S. citizen's entry. To account for the test rate, one may want to look at daily test positivity rate (daily new cases / daily tests) in OWID[5] and note that also here U.S. has relatively high rates compared to other European countries. The U.S. test positivity rate of about 8% is not really bad per se (Sweden had higher rate for considerable time), but may be seen as too high by some European countries, for the purpose of entry regulation.
From reading the L.A. Times article, I have an overall bad impression of the article; I would not rank the article as part of serious and neutral journalism. --Dan Polansky (talk) 12:10, 10 August 2020 (UTC)

Along with the different policies about who gets tested — anyone (U.S.), or those with symptoms (other countries) — there may be some other apparent policies which would help explain high U.S. case totals: nursing homes and contact tracing.

Nursing homes have made up around 50% of all reported deaths in the U.S. and Europe.[6] However, back in March, the U.S. Centers for Medicare & Medicaid Services (CMS) announced new policies focused on nursing homes, requiring testing nearly everyone, including workers, and reporting cases.CMS While in Europe they only recently began testing and reporting cases or deaths at nursing homes: "Public health officials around the world excluded nursing homes from their pandemic preparedness plans and omitted residents from the mathematical models used to guide their response."NY Times 8/8/20 That kind of variation in policies would likely explain why the cases count in the U.S. is relatively high.

Contact tracing, whether by a computer app or by personal choice, also seems quite different. In the U.S. the official CDC guideline followed by most states has been to test those "who have been in close contact with someone who is suspected to have COVID-19 or who has had a laboratory-confirmed positive test." Which explains why so many schools closed after even a small percentage of students or faculty tested positive. In Europe, however, while there are some new contact tracing apps being used, only about 15% of people use them. And the official policies still do not recommend testing contacts without symptoms or closing schools.[7] That would obviously produce more cases among U.S. students, although no symptoms are shown.

Taken together: excluding tests without first having symptoms, excluding nursing homes, and not testing contacts to those likely exposed, would imply that relative to the countries having those policies, the reported cases in the U.S. would naturally be much higher. Based on WHO's report that 80% of infections are asymptomatic, had the U.S. followed Europe's guidelines, wouldn't it therefore have about 1 million reported cases? --Light show (talk) 19:17, 10 August 2020 (UTC)

The above implied notion that European countries currently recommend against testing asymptomatics during contact tracing is bizarre and unsubstantiated (in response to "[...] the different policies about who gets tested — anyone (U.S.), or those with symptoms (other countries) — [...]". Asymptomatics are included in testing in multiple European countries and so do people in nursing homes. Moreover, Europe is not a single country and not a single federation; testing regimes vary among countries; and therefore, "Europe's guidelines" does not seem to refer to anything. The above contains other unsubstantiated claims that I will not comment on. To get information free from wild speculation, the reader is well advised to consult e.g. Our World in Data, where they can see for themselves what the testing rates are[8] and other relevant figures I linked above. --Dan Polansky (talk) 07:14, 11 August 2020 (UTC)
I don't think it was implied that European countries "recommended against testing" people without symptoms. For example, a recent BBC article states that "In England and Wales you can apply for a swab test for yourself, or for anyone in your household, if you or they have symptoms." That's also the guideline posted by Europe's ECDC. In any case, the link to testing figures seems to show that U.S. testing rates vary from 4 to 2 times the rates in Europe and elsewhere. The link also shows that as of today, in Europe, only Portugal, Germany and Greece, have an "Open public testing" policy for asymptomatic persons. So it doesn't appear that any of the cited statements above are based on "wild speculation." --Light show (talk) 18:35, 11 August 2020 (UTC)

Semi-protected edit request on 11 August 2020

  1. This phrase is missing a word: "but never progressed human trials due to a lack of funding" Please add "to" before "human"
  2. "may also be hobbled by long-standing distrust among minorities or public health officials" This has a typo, "or" instead of "of". The source prints it properly, so a Wikipedia editor made the mistake.
  3. "By April 25, The United States ranked 10th in deaths" Please decapitalize "The"
  4. "and demanded that their state" This is talking about protests in a number of states, so "state" should be pluralized
  5. "57% feel the federal government is not doing enough to address the limited availability of COVID-19 testing. 58% feel the federal government is not doing enough to prevent a second wave of COVID-19 cases later in 2020." This is a poll from three months ago, so "feel" and "is" should be in the past tense.

71.206.133.192 (talk) 22:58, 11 August 2020 (UTC)

  Done. #4 was done as "their respective states".  Ganbaruby! (Say hi!) 08:56, 12 August 2020 (UTC)

Daily charts are broken

The daily case/death charts, x axis, is now labelled only with months. But the labels such as April and May do not correspond to April 1, May 1, etc. This can be easily seen by looking at the charts from a few days ago, which had day-of-month numbers. 67.169.166.36 (talk) 09:40, 17 July 2020 (UTC)

I confirm the problem; the plots can be compared to those available at worldometers. I saw similar problems of bad x-axis labeling for x-axis of the date type in charts I was making on Wikiversity. It seems to happen once the number of data points exceeds some number. Perhaps someone would be inclined to open a ticket against the graphing add-in, but the issue may lie deeper, in the graphing library.
In the meantime, the issue could be addressed by a workaround: reduce the number of data points e.g. by dropping the starting values, in February and early March. When I dropped the first 10 values (Feb 26-Mar 6), the problem disappeared. I could drop that data but let me note that as new data points are added, more values will need to be dropped from the beginning. --Dan Polansky (talk) 07:13, 18 July 2020 (UTC)
I went ahead and dropped first 10 values from multiple charts to fix the above so they now start at Mar 7. It helped. The visual information loss seem tolerable. But it is not ideal. --Dan Polansky (talk) 07:34, 19 July 2020 (UTC)

The fix Dan made works (thank you!), but as he said it's only temporary. I am completely unfamiliar with the charting tools, so I suggest someone more experienced should open a bug ticket.

After a second cup of coffee, I now see that my original report duplicated that from TrilliumLady a few hours before me.

I also now see the same (?) bug is affecting other charts on this page, but less obviously. As of today (July 21) both the >100,000 and 50,000-100,000 cases charts are depicting data points out to about July 28 (7 days in the future).

Another Talk poster mentioned the obvious 7-day periodicity in the daily charts. It seems pretty clear that some people in the reporting chains have ordinary 5-day work weeks.

The daily charts were very informative back in March when things changed rapidly. Today, things change slowly. So maybe the daily chart for Feb-May could be made static, as an archive, and then a new current chart could show weekly averages? That would reduce the data point count (1/7x), avoiding (for now) the charting bug, and it would hide the artificial weekend lag too. 67.169.166.36 (talk) 10:37, 21 July 2020 (UTC)

I fixed the two charts in "Number of U.S. cases by date" section as well, and even anonymous IP users can do that. Having weekly numbers sounds interesting; I am not sure how comfortable the updaters would be with that idea since their data sources probably show daily values. (The updaters still did not tell us what their sources are.) --Dan Polansky (talk) 11:48, 21 July 2020 (UTC)
The plots would ideally be smoothed by applying 7-day moving average, and I could do that but I do not know how the updaters would cope with that. Calculating the average of 7 values is easy, but it is harder than copying a value from one location to another. Ideally, the plotting framework would allow something like "y2=sma(y1, 7)", and calculate that automatically; "sma" stands for "simple moving average". --Dan Polansky (talk) 12:02, 21 July 2020 (UTC)
The charts were significantly changed, in particular their size, which now makes their utility significantly less. I would recommend a larger size for the charts, especially those with the data of many states. Jaedglass (talk) 05:55, 24 July 2020 (UTC)
I set the width=700 from 900 previously (diff) because the charts were too wide and now they look perfectly fine on my 15" screen, not small at all, with no reduced utility. What screen size are you using or what kind of device that they seem too small to you? (In fact, I would be happy with width=500. And there are larger screens than 15" but also smaller screens.) --Dan Polansky (talk) 08:33, 25 July 2020 (UTC)
15.5 inch Dell Inspiron set with recommended resolution of 1366X768, and given the number of lines on the chart, at the current size the charts are not meaningful.Jaedglass (talk) 07:22, 26 July 2020 (UTC)
I got similar screen size and resolution as above. The charts and their level of detail appear meaningful. The number of horizontal lines appears sufficient. However, the number of vertical lines may be low if the chart starts showing month boundaries, which it does if the number of points exceeds a certain threshold, it seems; right now, the "Number of U.S. cases by date" charts have good number of vertical lines, which was ensured by someone by removing some data points at the beginning. And right now, e.g. the "No. of new daily cases" chart has very few vertical lines, which is fixed not by increasing the chart size but rather by dropping several initial data points. --Dan Polansky (talk) 17:14, 27 July 2020 (UTC)
I will clarify that I am most specifically discussing the multi-state charges the number and width of the lines given the height of the chart and similar trends makes comparing states difficult. The charts would not have to be vastly bigger to meaningfully improve readability. Single line charts do not have the same issue. Jaedglass (talk) 05:42, 31 July 2020 (UTC)
In "the number and width of the lines", you mean the colored data lines, not the grey grid lines, I guess. In the multi-data-line charts showing cumulative (total) cases for multiple states, I can easily compare states and see trends (up, down, plateau, linear, faster-than-linear). Let's try: name me two states and based on the charts I will give you a verbal comparison of the two states. For a start, in the upper chart we can see that California and Florida have the most cases in absolute terms (we do not see per capita) and that they are still growing relatively fast but linearly; we can see that California and Florida overtook New York rather recently and that New York was the leader of the pack by a wide margin for a long time. On another note, I increased the chart height now to 250. --Dan Polansky (talk) 08:56, 31 July 2020 (UTC)
Let me note that for many purposes, the two multi-state charts are not so useful since they do not show per capita cases but rather absolute cases. Thus, two states that have the same case-growth rate per capita but significantly different populations will be shown as having different slopes in the charts. The two charts would be much more useful if redone on a per capita basis, but to make it manageable for the updaters, probably a source would have to be found that provides daily cumulative per-capita cases already calculated. For an interstate comparison of case growth, test positivity rate (=daily new cases/daily new tests) is much more useful and is available at coronavirus.jhu.edu; they also have a colored map (click on "Map View") that shows new confirmed cases per 1,000 people in different states and for each state the number is available on a mouseover. --Dan Polansky (talk) 09:18, 31 July 2020 (UTC)
Yes, information about New York was clear because the numbers of its coronavirus cases, and therefore its line, were significantly separated from that of any other state. The same can be said now of California and Texas because of their rapid upward trajectory. However, other than those three states, my view is that little can be gleaned from the mass of lines of all the other states, and that a bigger chart would still be helpful. Whether absolute figures are meaningful or not is a separate question, but that seems like something that can be included in the 50 state table, if that information is not already included.Jaedglass (talk) 05:42, 4 August 2020 (UTC)
As for "little can be gleaned from the mass of lines of all the other states", that is too unspecific to be testable. For each state, I can see 1) its current approximate case value, 2) the recent rate of change (slope of the curve), 3) the relation of the two pieces of information to other states. Sometimes the lines meet in a cluster but even then I can see what is going on; to wit, e.g. Massachusetts cumulative cases are now at about 120 000 and are growing slowly, whereas cases for other states in the same cluster grow faster, which include Tennessee, Louisiana, North Carolina and more. --Dan Polansky (talk) 08:39, 4 August 2020 (UTC)
My only thought on what "little can be gleaned..." would be that the clumps get hard to identify colors that are hidden behind others. It's not perfect, but I agree with Dan Polansky, that it is useful even if not perfect. Scotty.tiberius (talk) 11:22, 13 August 2020 (UTC)

Charts

There are some problemes with the numbers for Louisiana and Massachusetts. Some days simply repeat the figure of the previous day, which doesn't seem correct. I could correct similar cases in TN and TX, but do not find the timeline for LA and MA. --Qumranhöhle (talk) 16:54, 6 August 2020 (UTC)

Yeah. **LOTS** of states don't update every day.  :-( TrilliumLady (talk) 06:21, 14 August 2020 (UTC)

Moving info from lede instead of deleting it

@Michelangelo1992: Regarding this removal...it's certainly a fair argument that this is too much detail for the lede. It would be better, though, to move material like that to the body rather than simply deleting it, if that's the only reason it shouldn't be there. (I did that for this case.) -- Beland (talk) 20:19, 15 August 2020 (UTC)

The removed material about hospitals having reached capacities is very interesting and relevant, and should probably be somewhere in the article in some form. However, care should be taken that it is neutral. For instance, when it says "Arizona declared crisis standards of care in July 2020, allowing hospitals to legally provide treatment normally considered substandard to some patients in order to save others", it should also says that Arizona never actually needed to make use of the crisis standards of care (that's as far as I know, and I'll stand corrected by references). --Dan Polansky (talk) 07:37, 16 August 2020 (UTC)

Simplistic death projection

A simplistic cumulative death projection follows.

We take the U.S. population to be 330,070,331 per List of countries and dependencies by population. Statista.com instead assumes 327.17 million[9].

Current U.S. cumulative deaths are 155,224[10].

Current U.S. cumulative deaths per million pop is 474.45[11].

If we assume the ultimate death rate will reach the U.K.'s 695[12], multiplying by population and rounding to thousands gives us:

230 000 deaths for the U.S.

If we assume the ultimate death rate will reach current worst rate, Belgium's 862.37[13], we get:

285 000 deaths for the U.S.

If we get a bit more optimistic and assume Spain's rate 609.37[14], we get:

201 000 deaths for the U.S.

The above is very approximate and simiplistic, but gives the scale of things. --Dan Polansky (talk) 09:19, 4 August 2020 (UTC)

  • Well, according to an estimate by Michael Osterholm made in the end of April, there is a possibility of at least 800,000 deaths in the USA over the next 18 months [15]. But even a more important question is what will happen with health of people who survived the disease [16]. My very best wishes (talk) 04:21, 6 August 2020 (UTC)
    Given what we know now, 800,000 final deaths for the U.S. seem extremely unlikely: that would correspond to death rate per million pop of 2,424, nearly triple of Belgium and much higher than New Jersey's 1,792[17]. At the end of April, we knew much less, especially we did not know the Swedish development until the beginning of August and the development in other European countries. We did not know there is probably a T-cell-based immunity in addition to antibody-based immunity.
    And tracing an estimate to a person rather than a method of estimation published in an article is fundamentally unscientific: if estimate E is traced to article A, we can critically examine article A; if estimate E is only traced to person P, we can only critically examine the person, so we leave the domain examined (epidemiology) and enter the domain of sociology or credentiology of science. And the credentiology does not get us very far since scientists are fallible, and even some of the most skeptical and careful scientists are prone to error. While Michael Osterholm is an epidemiologist, he would have to publish an article or statement; the first source given above is an opinion piece by CNN, which is fundamentally unreliable for science. --Dan Polansky (talk) 09:20, 6 August 2020 (UTC)
  • "seem extremely unlikely" Said who? You? This is not "an opinion piece by CNN". 800,000 of deaths is a highly approximate estimate by Michael Osterholm, and he explained how did he came to such number in the cited RS. As soon as an estimate by a notable expert has been reliably published (yes, it was reliably published), it belongs to the page. Is it "the truth"? No, of course not, just as all other estimates. But there is a significant range of estimates by various research teams and individual researchers who are experts. My very best wishes (talk) 16:35, 6 August 2020 (UTC)
Personally, I think his estimate may or will be proven wrong if the new vaccine will be introduced much sooner than he thought. Yes, it looks like it might be introduced very soon. My very best wishes (talk) 16:44, 6 August 2020 (UTC)
https://www.cnn.com/2020/04/21/opinions/bergen-osterholm-interview-two-opinion/index.html is an article in CNN in the category "opinions", so it is an opinion piece. Scientists do not publish science in CNN; they publish science in peer reviewed journals or they may have to resort to a preprint. The 800,000 deaths are very unlikely based on the statement provided, especially the corresponding death rate per million pop. The unlikeliness does not depend on any vaccine. The very approximate projections I made above do not assume any vaccine; they are made with the advantage of the hindsight available on Aug 4, that is, the advantage of death rates per million in countries that are likely to have already seen great majority of their covid deaths. --Dan Polansky (talk) 16:57, 6 August 2020 (UTC)
As for Michael Osterholm, I wonder whether anyone can point out any peer-reviewed scientific articles where he is an author, possibly one of multiple authors; his Wikipedia page does not help in that regard. --Dan Polansky (talk) 17:08, 6 August 2020 (UTC)
And his estimate from the quoted CNN opinion piece works as follows: "[...] About 0.5 to 1% of the total number of 160 million infected people will die. So you have the possibility of at least 800,000 deaths in the US over the next 18 months." Thus, to arrive at the 800,000 deaths, he assumes 50% of the population will be infected (160 million) and of them 0.5% will die. Using the same assumptions--50% infected, 0.5% die--for Sweden, we get 25,575 deaths for Sweden given 10.23 million pop. But we have only seen about 5,000 deaths and they appear do be done already. So his assumptions seem suspect. --Dan Polansky (talk) 17:22, 6 August 2020 (UTC)
As for "As soon as an estimate by a notable expert has been reliably published (yes, it was reliably published), it belongs to the page": that really caught my attention. Some people seem to think that science includes durably recorded utterances of notable experts; some politicians say "we followed science" but what they really mean is "we followed our officially appointed experts and those experts who agreed with them". Expert opinion is not identical to science, especially peer-reviewed science. Bowing to authority, especially expert authority, is antithetical to scientific spirit proper. --Dan Polansky (talk) 17:36, 6 August 2020 (UTC)
This is a reliably published estimate by a well known expert in the field, hence a perfectly valid content per WP:RS and WP:NPOV. He is not a politician, but an expert. It was an estimate for 18 months assuming no vaccine or effective medications being developed and widely used during this time. And no, Sweden is not "done already" with the pandemic. None of the countries is "done". My very best wishes (talk) 17:45, 6 August 2020 (UTC)
My position is that "reliably published estimate by a well known expert in the field" is no science and that a CNN opinion piece is not reliable source on science. Anyone can see above how simplistic the estimate was, and that it was no more complicated, sofisticated, informed or scientific than the projections that I posted above. Sweden does appear to be done with the deaths since its daily all-cause deaths are now below the reference average, but let us see; I stand by the notion that multiple other European countries such as UK and Spain are likely to have already seen great majority of their covid deaths, but absolutely certain it is not and let us see. --Dan Polansky (talk) 17:55, 6 August 2020 (UTC)
Sure, not everything that a scientist said was science. But why science? If you look at the sources currently used on this page, a lot of them are not science, but they are RS and therefore consistent with all policies, such as WP:RS. My very best wishes (talk) 18:20, 6 August 2020 (UTC)

Other estimates, from tweet photo by Michael Levitt, chemistry nobel laureate, biophysicist, May 27:

Estimated U.S. deaths based on the scaling factor to match Diamond Princess Deaths:

264,081

Estimated U.S. deaths based on the scaling factor to match Wuhan Deaths:

203,837

Click on the photo in the tweet (screenshot of Excel sheet) to see these estimates. The scale of these estimates is the same as those projections I made above, except that I made them of Aug 4 in a very simplistic way using hindsight, whereas Michael Levitt's estimates were made on May 27 with much more sophisticated method. --Dan Polansky (talk) 10:29, 16 August 2020 (UTC)

As deadly as the 1918 flu pandemic - not

In the news:

An interesting quote from the above:

'Dr. Anthony Fauci, the nation’s leading infectious disease expert, has said the coronavirus is a “pandemic of historic proportions” and history books will likely compare it to 1918.'

News reseach suggests Fauci indicated the above on Jul 15, 2020 or thereabouts.

Let's have a look:

All-cause deaths in Sweden in Oct–May, calculated from SCB [18] [19], for 1900–2020:

Above, each year on the x-axis is the year of Jan–May data, while Oct–Dec data are for the previous year. Beware that the above is not adjusted for population.

The pandemic seems to be largerly over in Sweden. The conclusion is obvious. --Dan Polansky (talk) 15:02, 15 August 2020 (UTC)

The Fauci quote does not appear in this article, and this article is about the United States, not Sweden. COVID-19 pandemic in the United States#Death projections has an apples-for-apples comparison of COVID-19 and the 1918-19 flu epidemic. -- Beland (talk) 20:23, 15 August 2020 (UTC)
The mentioned section provides a similar conclusion as the above chart. However, the section is based on CDC estimates of 1918-19 flu deaths, while the above chart shows the statistical actuals, albeit for Sweden; and the epidemic is not entirely over in the U.S. yet. Admittedly, the statistical actuals are from a different country, and if severity of the 1918-19 flu differed widely between the U.S. and Sweden, it would not be relevant for the U.S. But if we assume that the severity of 1918-19 flu was similar between countries, or if we assume that at least we have no reason to believe it was milder in the U.S. than in Sweden, we get that the chart is very relevant; we can also note that the current per million death rate of U.S. is still smaller than Sweden's[20]. The consideration I just made suggests the chart is very revealing for U.S. also.
On another note, Fauci is currently quoted by this article as if he were a reliable expert; the present section shows otherwise. --Dan Polansky (talk) 07:03, 16 August 2020 (UTC)
He is, you aren't, you are just a wikipedia user. And your above section is a violation of the basic wikipedia principle WP:NOR. --Qumranhöhle (talk) 07:45, 17 August 2020 (UTC)
The reader perusing what I posted above can conclude, based on the information presented rathen than its author, that Fauci is not a reliable expert. A computing-literate user can verify the chart; it does not require programming. Let's avoid the ad hominem fallacy that tries to infer from properties of a person (e.g. "just a wikipedia user") the lack of veracity of arguments and proofs produced by the person. Rather, the proofs, arguments and evidence need to be examined. (I do not deny that Fauci is a nominal expert: such can be easily verified in the media. He is not a real expert.) --Dan Polansky (talk) 08:11, 17 August 2020 (UTC)
Your ad hominem-argumentation against an acknowledged expert is not appreciated. You did not and in fact cannot present any claim that Dr. Fauci is not a "real expert", it is just your so called (baseless) "opinion", which is irrelevant here. Stating that you are "just a wikipedia user" is a fact, that puts you in your place. (Calling you a Trump troll and Putin Troll or a Covidiot etc. would be an ad hominem argument, but I didn't do so.) Dr. Fauci did not - in that article - say, as you suggest, that this pandemic is "as deadly as the 1918 flu pandemic". (Yet it is deadly enough and there is no end visibl yet, so better focus on fighting the pandemic instead of applauding to corrupt political clowns that pretend to be president.)
Furthermore, it can be easily shown that your claim is absolutely baseless. Dr. Fauci is quoted with "pandemic of historic proportions", which only few Covidiots or conspiracy theory freaks will deny durrently. The article furthermore summarizes his words with "and history books will likely compare it to 1918". This is, by the way, exactly what you did: comparing this pandemic with 1918. But again, this is irrelevant as per basic wikipedia principles and it is furthermore irrelevant for this article. --Qumranhöhle (talk) 10:29, 17 August 2020 (UTC)
Let the reader read what I posted and make their own judgment; I have no patience with the low-grade substance and the bad tone of the above. --Dan Polansky (talk) 11:12, 17 August 2020 (UTC)
As I have little patience with trolls, liars or wikipedia users who think they know better than acknowledged experts. --Qumranhöhle (talk) 11:21, 17 August 2020 (UTC)
The above is very rude, but a key thing is that if we play the credentialism game, chemisty nobel laureate Michael Levitt has much better credentials than Fauci, and is a higher card to play. --Dan Polansky (talk) 11:25, 17 August 2020 (UTC)
For the reader, I will provide direct evidence:
I transcribe one sentence (please double check): "it's something that I think when history looks back on it it will be comparable to what we saw in 1918"
As I understand it, "comparable" above means "of similar impact in terms of death". In that quote, "comparable" does not mean, "such that the severities are capable of numerical comparison", since if it would mean that, the covid pandemic would be "comparable" to any other pandemic, and that cannot be the intended meaning. --Dan Polansky (talk) 11:25, 17 August 2020 (UTC)
Of "historic proportions" it is, comparable in its scale of death per capita to 1968 flu pandemic for some of the worst affected regions. To claim in July that it will be rather comparable to the 1918 flu pandemic is patently absurd. --Dan Polansky (talk) 12:11, 17 August 2020 (UTC)
You played the card wikipedia user Dan Polansky vs. Dr. Fauci, that's patently absurd. And with all due respect to nobel laureate Michael Levitt, several of his "predictions" about COVID19 have already been proven wrong. Despite that, I would never dare to say this undermines his general credibility, it simply shows he was wrong about COVID19, in other words: yes, even experts can be wrong - and Levitt was wrong here, so it was not helpful for your case to "play this card". In any case that doesn't make random wikipedia users "real experts". --Qumranhöhle (talk) 12:27, 17 August 2020 (UTC)
I did not play that card since I originally did not play the credentialism card at all: I presented an argument and evidence, not my credentials. I believe the arguments and evidence presented can be understood by the general reasonably well educated reader. As for the several predictions being wrong, please share with us some specific examples so that we know what we are talking about and how much he was wrong. To dismiss my arguments only because they are mine is what was done above, and what I think is subpar. The July (not March) comparison to the 1918 flu is a huge red flag, I believe, but others may differ. --Dan Polansky (talk) 13:10, 17 August 2020 (UTC)
No, you are misusing an article discussion page to badmouth an acknowledged expert (despicably continuing the trash coming from the White House), trying to call his expertise into question. Fact is, you as an ordinary wikipedia user are simply not qualified to make such statements. Nobody is interested in your credentials, they don't count here. If you are an expert, then write something in the real world and look what other experts say.
Another fact, you introduced Michael Levitt here and wanted to make a point. Somehow that fired back to you. Your problem, not mine. --Qumranhöhle (talk) 13:25, 17 August 2020 (UTC)
P.S.: "I did not play that card since I originally did not play the credentialism card at all" - does that really make sense to you? OMG! --Qumranhöhle (talk) 13:27, 17 August 2020 (UTC)
Share with us some specific examples. --Dan Polansky (talk) 13:30, 17 August 2020 (UTC)
I assumed you are able to look up the wikipedia article yourself. --Qumranhöhle (talk) 13:46, 17 August 2020 (UTC)
Please provide specific quotes that you had in mind. --Dan Polansky (talk) 13:54, 17 August 2020 (UTC)
So you are not even able to read the wikipedia article and check the links there? Thanks, I've seen enough. Back to the beginning: Your badmouthing of Fauci is just despicable and your own position is irrelevant. Please stop it, it embarassingly destroys the credibility of your account even further. --Qumranhöhle (talk) 14:03, 17 August 2020 (UTC)
(Outdent) The allegation made is this: 'And with all due respect to nobel laureate Michael Levitt, several of his "predictions" about COVID19 have already been proven wrong.' I request substantiation of the allegation in the form of specific quotes by Michael Levitt, as is standard with such requests. --Dan Polansky (talk) 14:08, 17 August 2020 (UTC)
And I#ve referred you to the wikipedia article which you are unwilling to read. For the willing readers: https://www.jpost.com/israel-news/nobel-laureate-israel-will-have-no-more-than-ten-coronavirus-deaths-621407 from March 20. How many COVID19-dead does Israel count at the moment? From the same interview: "Italy is already half way through the disease". How many dead had Italy at that time, how many does Italy count at the moment? Enough? --Qumranhöhle (talk) 14:13, 17 August 2020 (UTC)
Thank you; the quote from the article is "I will be surprised if the number of deaths in Israel surpasses 10". He was wrong about that; the covid-coded deaths in Israel are now 690 per Worldometers[21]; I would need to look at excess deaths in Israel, though; maybe later. Let me note that it was in March, rather early into the pandemic, but still. How can anyone predict the pandemic will be comparable to the 1918 flu in mid July? Mind boggles. --Dan Polansky (talk) 14:23, 17 August 2020 (UTC)
And the most obvious question at all: What does all of the above tell us about credentialism, about blind and uncritical reliance on experts? --Dan Polansky (talk) 14:24, 17 August 2020 (UTC)
As for Israel deaths, covid-coded deaths are not the same as covid-caused deaths. Wikiversity:COVID-19/All-cause deaths/Israel gives an interesting picture. --Dan Polansky (talk) 14:30, 17 August 2020 (UTC)
Do you have no decency at all? --Qumranhöhle (talk) 15:48, 17 August 2020 (UTC)
The problem with personal attacks like the one above is not only that they spoil the discussion atmosphere but also that they are very uninformative. What we do not learn from the above is: 1) which of my statements does the above editor find very problematic, and 2) which evidence or deliberation can be supplied to refute or weaken my statements. The reader learns very little, if anything at all. --Dan Polansky (talk) 16:39, 17 August 2020 (UTC)
The discussion atmosphere is spoiled by users who try to disgrace acknowledged experts and instead present themselves as experts while violating basic principles of wikipedia. The discussion atmosphere is also spoiled by users who prefer to deflect when they are disproved and start a new topic instead. Such users are not capable of giving satisfaction and just a disgrace to wikipedia. --Qumranhöhle (talk) 19:14, 17 August 2020 (UTC)
What exactly are you proposing? Wikipedia is WP:NOTAFORUM. Please use the discussion board to make suggestions to the article instead of just discussing the pandemic in general. Michelangelo1992 (talk) 19:30, 17 August 2020 (UTC)
This section was to add a little piece of evidence that experts cannot be uncritically relied on and should be contrasted to peer-reviewed science and to available data, in support of the proposal to remove or curtail Requests for a Reset section made in #Requests for a Reset Section on this talk page. Meanwhile, the section was removed in diff. The removal was done by SusanLesch, with whom I had a reasonably amicable discussion in #Requests for a Reset Section, which involved exchange of substance and avoided any possible revert wars and personal attacks; thank you to SusanLesch. --Dan Polansky (talk) 06:28, 18 August 2020 (UTC)

Plotting test positivity rate

The test positivity rate plotting syntax is produced by the following Python script ("plotTestPosRate.py"):

import sys, csv

fileName = sys.argv[1] # owid-covid-data.csv
countryCode = sys.argv[2] # e.g. USA
dayCountForAvg = 7

series = []
with open(fileName) as file1:
  for line in csv.DictReader(file1):
    if line["iso_code"] == countryCode:
      newTestsStr = line["new_tests"]
      if newTestsStr != "":
        testPosRatePerc = 100 * float(line["new_cases"]) / float(newTestsStr)
        series.append( (line["date"], testPosRatePerc) )

idx = 0
avgSeries = []
currentSum = 0
for date, testPosRatePerc in series:
  currentSum += testPosRatePerc
  if idx >= dayCountForAvg:
    currentSum -= series[idx - dayCountForAvg][1]
  if idx >= dayCountForAvg - 1:
    avgSeries.append( (date, currentSum / dayCountForAvg) )
  idx += 1

datesOut = ", ".join([k for k, v in avgSeries])
ratesOut = ", ".join(["%.1f" % v for k, v in avgSeries])
sys.stdout.write("|x = " + datesOut + "\n")
sys.stdout.write("|y = " + ratesOut + "\n")

Usage:

plotTestPosRate.py owid-covid-data.csv USA

Other country codes are possible.

File owid-covid-data.csv can be obtained from OWID github[22].

The output of the script can also be obtained more manually via a spreadsheet.

--Dan Polansky (talk) 08:08, 18 August 2020 (UTC)

Most testing per capita

The lead of our article currents claims that "As of mid-August, the United States had the most per capita testing of any other country", sourced to JHU. However, all the data I've checked shows that New Zealand overtook the US in per capita testing on 13 August (3.26 vs 2.26 tests per thousand) and has remained higher since (16 August: 5.39 vs 2.35). See https://ourworldindata.org/coronavirus for example. I don't think we can leave patently untrue claims in the article, especially in the lead. --RexxS (talk) 18:28, 17 August 2020 (UTC)

The JHU source seems to indicate different numbers. Frankly, I am not sure which of them is correct (if either). I am fine with removing the contested sentence barring clearer sourcing. Thank you for pointing this out. Michelangelo1992 (talk) 19:32, 17 August 2020 (UTC)
On the other hand, it would be necessary to explain why flooding the article with charts and graphs from JHU figures is fine, but one selected figure is not. And comparing New Zealand to the U.S. is not a valuable comparison, since its entire population is the size of a large American city.--Light show (talk) 19:39, 17 August 2020 (UTC)
I don't understand your disagreement here. Are you suggesting leaving the article as-is? As RexxS pointed out, sources seem to disagree about which country has the highest test positivity rate; I am not sure which is correct. Whether or not the USA is "comparable" to New Zealand is irrelevant. When making statements about the per-capita testing rate of different countries, it is obvious that sources must compare countries regardless of population density or other country-specific factors. Additionally, you reverted my edit on the lede sentence about test positivity rate, despite the fact that the sourced JHU graph showing the WHO recommended guideline of 5% and the USA test positivity rate >7%. I have added another source to make this more clear. Michelangelo1992 (talk) 20:21, 17 August 2020 (UTC)
Light show, how does the population of New Zealand matter when the metric is per-capita testing, not raw numbers of tests? 64.203.187.119 (talk) 20:28, 17 August 2020 (UTC)
Considering a simple fact such as the U.S. having 10 cities with larger populations than the 10 smallest countries in Europe, to claim that a detail like that is irrelevant is hard to rationalize. Also, I didn't revert your edit, I split your (still) unsourced comment to a separate sentence. The way it was thrown in there inline without a source or even supported commentary in the body, gave it the effect of being a "yeah, but..." comment. The way you sourced it is essentially a synth, forcing readers to look at charts to support the statement. With a massive article like this one, a lead statement should be clear from the body text, which it is not. --Light show (talk) 20:55, 17 August 2020 (UTC)
I reorganized the sentences. I also added an NYT source. Michelangelo1992 (talk) 21:09, 17 August 2020 (UTC)
@Light show: "And comparing New Zealand to the U.S. is not a valuable comparison, since its entire population is the size of a large American city" - NZ has an estimated population of about 5 million people. Tell me which US American city has 5 million inhabitants, apart from New York? --Qumranhöhle (talk) 21:24, 17 August 2020 (UTC) P.S.: "Considering a simple fact such as the U.S. having 10 cities with larger populations than the 10 smallest countries in Europe" - the question was about New Zealand which is not part of Europe, just to be sure... --Qumranhöhle (talk) 21:25, 17 August 2020 (UTC)
List of North American metropolitan areas by population, since most of the cities are considered metro areas which tend to be split up. --Light show (talk) 21:42, 17 August 2020 (UTC)
So you were inexact, speaking about cities but referring to metropolitan areas. --Qumranhöhle (talk) 21:53, 17 August 2020 (UTC)
My original point wasn't that the article was claiming "As of mid-August, the United States had the most per capita testing of any other country, except for some that aren't very big, even though they are countries"   --RexxS (talk) 22:47, 17 August 2020 (UTC)
Yes, that's a problem per OWID. One solution is to weaken the statement to say "has one of the highest per capita testing rates among countries" or the like, which is true, verifiable in OWID, and probably reasonably stable in time. I don't oppose removal of the statement either. --Dan Polansky (talk) 06:42, 18 August 2020 (UTC)
The solution should be to comply with WP:BALANCE:

when reputable sources contradict one another and are relatively equal in prominence, describe both points of view and work for balance.

There's a detailed explanation at WP:ASSERT. I suggest something along the lines of this:
  • As of mid-August, according to Johns Hopkins University, the U.S. has been doing the most per-capita testing of any country,{JHU ref} although Our World in Data shows New Zealand having a higher rate since August 13.{OWID ref}
The OWID ref isn't easy to give directly, because you either have to download the entire dataset and look though it, or use the interactive graphing at https://ourworldindata.org/coronavirus. It's also possible to look at https://ourworldindata.org/coronavirus-testing#new-zealand and https://ourworldindata.org/coronavirus-testing#united-states (two sections of the same page), but none of them neatly fit our fixed url format for a source. --RexxS (talk) 21:27, 18 August 2020 (UTC)
Whatever. It seems like an overkill given this is not a sport game; being the top vs. being among the top by testing rate does not make material difference. Better pick some countries and show their testing rates per capita in one graph, but I do not know how to pick the countries. Maybe pick top 7 countries by testing rate per capita and plot them? --Dan Polansky (talk) 08:54, 19 August 2020 (UTC)
I added a chart, with intro phrase starting with "Daily new tests per 1000 pop [...]". The code is below in #Plotting tests per 1000, as a comparison. --Dan Polansky (talk) 14:14, 19 August 2020 (UTC)
The charts are nice, but surely it's not overkill to ask for accuracy in the lead? This isn't a question of sports games, it's a question of our article making a claim in the lead that is obviously contradicted by another reliable source. Never mind, I'll do it myself. --RexxS (talk) 20:18, 19 August 2020 (UTC)
Hong Kong appears to test more than the U.S. according to Johns Hopkins. I'm not very good at reading these charts so maybe somebody else would check please? -SusanLesch (talk) 23:38, 19 August 2020 (UTC)
Hong Kong isn't a country, despite being bigger than some, so it doesn't contradict the claims made. Nevertheless, New Zealand, Israel and the UK are currently doing more daily testing per capita, according to OWID, and there may be others. --RexxS (talk) 00:07, 20 August 2020 (UTC)
I just looked at the sources supporting the JHU claim: International Comparison of Positivity Rates and Tests Per Capita and their dashboard. The latter gives no indication about per capita testing, so I've removed it. The former makes the claim "The U.S. has conducted more COVID-19 tests than any other country", but surely that's the cumulative figure? It's interesting to note that JHU sources its international testing data from OWID. Perhaps a re-write is indicated? --RexxS (talk) 00:31, 20 August 2020 (UTC)

(Outdent) The U.S. does not lead by daily new tests per thousand as per the chart in the article, and it also does not lead by cumulative (total) tests per thousand, smoothed (showing top 3 + USA per OWID data):

U.S. does not lead by the daily new tests, smoothed, absolute (showing top 3 per OWID data):

U.S. does lead by the total tests, absolute (showing top 3 per OWID data):

Absolute death counts and absolute test counts are uninteresting: they largely depend on how populous a country is. --Dan Polansky (talk) 10:12, 20 August 2020 (UTC)

Dropping CFR chart

I am about to drop the CFR chart. The chart is very misleading: it disregards different testing rates, and states with higher testing rates will discover disproportionally more asymptomatic and mild cases, which will lead to lower case fatality rate (CFR) via inflated CFR denominator. And the disparities between tests per million pop are significant: New York state has about 378 000 total tests per million pop while Vermont has about 185 000 total tests per million pop. Connecticut has about 284 000 tests per million pop, which explains why its CFR is higher than New York state's. I calculated the test rates from realclearpolitics.com data[23]. In general, CFR is a very unreliable indicator of anything.

The chart, for reference:

--Dan Polansky (talk) 09:08, 21 August 2020 (UTC)

Criticism of the PIRG letter

The PIRG letter web page: https://uspirg.org/resources/usp/shut-down-start-over-do-it-right; the PIRG letter pdf: https://uspirg.org/sites/pirg/files/USP_Public-health_final-letter-shutdowns_V4.pdf. From the pdf: "Of all the nations in the world, we’ve had the most deaths from COVID-19": starting the letter with sophistry consisting in comparing death absolute numbers rather than death rates per million pop is suspect. The letter does not contain any solid analysis, and there is no references section in the letter, although there is a link to https://www.statnews.com/2020/06/19/faster-response-prevented-most-us-covid-19-deaths/, a non-solid piece of non-science that I criticized in Talk:COVID-19 pandemic in the United States/Archive 13#Statnews article.

More from the letter:

  • "Right now we are on a path to lose more than 200,000 American lives by November 1st": true. Over 170,000 lives were already lost.
  • "And reopening before suppressing the virus isn’t going to help the economy": blatantly false.

Conspicuously absent is any mention of negative economic and health impact of the lockdowns, on local and global level. No mention of lives already lost due to lockdown can be found, or lives that are yet to be lost as a consequence of lockdowns.

Would anyone know of a reliable source criticizing the PIRG letter? If so, we might expand the article with criticism of the letter or at least a statement that criticism exists and a link to it. The criticism should be fairly obvious to a scientifically educated reader understanding the importance of peer-reviewed science and proper statistical analysis accounting for confounding factors. --Dan Polansky (talk) 08:16, 19 August 2020 (UTC)

Alternatively, I am wondering whether we could drop the paragraph mentioning the PIRG letter. While the letter is covered by partisan news outlets, I am not sure this is enough for inclusion in an encyclopedia; the outlets obviously did not bother do make a cursory critical review of the content of the letter. Right now, there is still an impression of bias in the article. --Dan Polansky (talk) 08:42, 19 August 2020 (UTC)

Nope. You wrote in Talk:COVID-19_pandemic_in_the_United_States#Requests_for_a_Reset_Section, "2) Dropping the paragraph altogether and keeping only the sentence about "[...] U.S. PIRG and 150 health professionals sent a letter asking [...]" would do Wikipedia a service." -SusanLesch (talk) 20:47, 20 August 2020 (UTC)
At that point, I had not read the PIRG letter and I had not realized how bad the material in the letter is. Mentioning the PIRG letter without mentioning how bad it is is biased. And the article does not mention any lockdown-opposing medical doctors (MDs); two opposing MDs had their video taken off YouTube. No attempt is made to determine how many MDs actually oppose further lockdowns. As it is, the portion of the article is biased. --Dan Polansky (talk) 09:07, 21 August 2020 (UTC)
And the PIRG letter even contradicts iself: it says "Listen to the experts" and yet there are all these medical doctors signed, who are not economic experts, at least as per their credentials, and thereby lack the expertise to sign the economic statement: "And reopening before suppressing the virus isn’t going to help the economy". --Dan Polansky (talk) 10:21, 21 August 2020 (UTC)

Semi-protected edit request on 17 August 2020

By July 9, the number of cases had passed three million.[146] President Trump...150,000 deaths.[149] On August 9, the U.S. passed five million COVID-19 cases.[150]

Could you merge the first and last sentences ("By July 9...three million,[146] and it passed five million on August 9.[150] President Trump...") and leave the remaining sentences as they are, following both numbers? It would be out of chronological order, but a short paragraph with numbers at both ends raises the question of "why are they separate?" It would read better if the rising-number-of-cases sentences were put together. 64.203.187.119 (talk) 20:26, 17 August 2020 (UTC)

How exactly are you thinking of wording the sentences? —Tenryuu 🐲 ( 💬 • 📝 ) 20:50, 17 August 2020 (UTC)
Why don't we just drop all the sentences of the form "confirmed cases passed X million?" I do not see any informative value in them. This is not a sport game. And they are misleading in so far as they disregard increased test rates; daily new tests were growing for multiple months. And we have charts for raise of cases. --Dan Polansky (talk) 06:59, 18 August 2020 (UTC)
This seems   Unnecessary. Closing. P,TO 19104 (talk) (contribs) 14:40, 21 August 2020 (UTC)

"Wikipedia:COVIDUS" listed at Redirects for discussion

  A discussion is taking place to address the redirect Wikipedia:COVIDUS. The discussion will occur at Wikipedia:Redirects for discussion/Log/2020 August 21#Wikipedia:COVIDUS until a consensus is reached, and readers of this page are welcome to contribute to the discussion. Hog Farm Bacon 14:45, 21 August 2020 (UTC)

"States affected with coronavirus" listed at Redirects for discussion

  A discussion is taking place to address the redirect States affected with coronavirus. The discussion will occur at Wikipedia:Redirects for discussion/Log/2020 August 21#States affected with coronavirus until a consensus is reached, and readers of this page are welcome to contribute to the discussion. Hog Farm Bacon 14:46, 21 August 2020 (UTC)

"Draft:COVID-19 pandemic in United States" listed at Redirects for discussion

  A discussion is taking place to address the redirect Draft:COVID-19 pandemic in United States. The discussion will occur at Wikipedia:Redirects for discussion/Log/2020 August 21#Draft:COVID-19 pandemic in United States until a consensus is reached, and readers of this page are welcome to contribute to the discussion. Hog Farm Bacon 14:48, 21 August 2020 (UTC)

"Draft:Coronavirus 2019 USA" listed at Redirects for discussion

  A discussion is taking place to address the redirect Draft:Coronavirus 2019 USA. The discussion will occur at Wikipedia:Redirects for discussion/Log/2020 August 21#Draft:Coronavirus 2019 USA until a consensus is reached, and readers of this page are welcome to contribute to the discussion. Hog Farm Bacon 14:55, 21 August 2020 (UTC)

Graphs in Number of U.S. cases by date section

I noticed that the y axes for the first two graphs in this section were re-calibrated a short time ago, which was definitely needed. However, I suggest making the the first one for states with >250K cases and the second one for states with 100K to 250K cases. With the split at 400K, Georgia is going to remain on the second graph for a while but keep compressing it, leading to the same problem as before, too many state curves overlapping and plots that will be tough to discern. If Georgia moves to the top graph when it passes 250K, that won't be a problem. Just my suggestion... Ira Ira Leviton (talk)

Generally agreed, but why not 200k? --Qumranhöhle (talk) 07:25, 19 August 2020 (UTC)
I see that 200K has already been done – it looks fine, and will keep the lines separated for a while longer (I hope). Thanks. Ira Leviton (talk) 17:10, 19 August 2020 (UTC)
What about the 9 "red zone" states (as of Aug 23, 2020) which are the hardest hit states in number of cases, per capita infection and positivity rates, oddly California isn't included on the list? They are Arizona, Arkansas, Florida (2nd highest number of cases), Georgia, Louisiana (the highest per capita infected), Mississippi, Nevada, South Carolina and Texas. The Oklahoma state governor Kevin Stitt had a meeting with Dr. Deborah Birx who considered OK a "red zone" state though it's not included as one, but he rejected her CDC-backed recommendations to curb the spread of COVID-19 in his state which had a 2-month spike that's probably the most of any in the USA during the pandemic. 2605:E000:100D:C571:E0A0:848E:577F:F036 (talk) 04:08, 23 August 2020 (UTC)