I did statistical analysis of seniority for some of the unions involved in the merger of Air Canada and Canadian Airlines 20 years ago. The stats consultants on the other side were very much into finding images in the charts, like "sailboats" that they could build a narrative around. I just looked at the charts to see if there was any obvious discrepancies. There were.
In the chart you show, what catches my eye immediately is the totally anomalous drop of non-COVID deaths below the orange line IMMEDIATELY after the peak of COVID deaths in roughly February(?) of 2021. As a student of mortality statistical probability going back to John Graunt in the 17th century, I can assure you that didn't happen. Notice that in the earlier two peaks, the "non-COVID" excess deaths rose in tandem with the COVID deaths. This time, they inexplicably dropped right after the peak? Nope. Not buying it.
What I think you have there is progressive accumulation of reporting delays. In other words, "noise." Unless you can explain why non-COVID deaths dropped in the wake of a COVID peak, statistical noise is the most reasonable explanation.
1. Death certificates "do not go into the data." Fifty-seven reporting jurisdictions report data to the National Center for Health Statistics based on death certificates. I doubt NCHS ever sees the death certificates. "Late filing does not create noise" is not what the NCHS statement about data limitations says. The data are weighted with estimates to correct for underreporting but those estimates cannot account for variability in cause of death.
2. It is noted that you didn't even acknowledge, let alone try to explain the SPECIFIC anomaly I pointed out, which is about the same size as the anomaly you are hawking. In my experience with statistics, cherry picking data is not productive of good analysis. I referred to above and below the orange line (Upper Bound Threshold) because that is what you showed in your chart. I am not at all surprised that Non-COVID deaths would be elevated above "Expected Deaths" because during a prolonged pandemic, many people postpone or are denied medical treatment that in normal times might prevent a future death. Because "Expected Deaths" is estimated based on non-pandemic conditions, it doesn't literally mean what would be expected NOW.
3. My "unintended" compliment was actually an intended one. I read and took seriously your deductions from the data and then considered whether there might be limitations to the data that would render your deductions as more like speculation. I would be very grateful if readers gave the same degree of consideration to my observations.
Fair enough. There is nothing wrong with spotting a possible emerging trend and calling attention to it. I don't think you have jumped to conclusions but have mooted several possible explanations, all of which are at least plausible. I would add, though, noise in the data as another plausible explanation. I wouldn't be shocked about silence from the CDC. Anything they say -- literally anything -- is subject to distortion, manipulation, and propagandizing. If they said the data is too incomplete they would be accused of concealing the facts. etc., etc., etc. This is not to say institutions do not conceal information that puts them in a bad light. I do have several epidemiologists in my immediate circle of friends who are notoriously critical of Big Pharma and consequently skeptical of the actions of the government bodies that regulate drugs. On balance, I've found the official pronouncements more credible than the alternative views on this one. Given the choice between the CDC and an "Anarcho-Libertarian" propaganda website, I would go with the former. I view right-wing libertarianism as a gateway drug to white supremacism.
"'What will people do if they are thrown out of work because of new technology, immigrants, or free trade?' It’s a common question that, too often, rests on the “lump of labor” fallacy; namely, the idea that there is only so much work to be done and no more, and that if one person takes a job then someone else is necessarily out of luck."
chill sandwichman no need to write a whole school essay on comment dog
I did statistical analysis of seniority for some of the unions involved in the merger of Air Canada and Canadian Airlines 20 years ago. The stats consultants on the other side were very much into finding images in the charts, like "sailboats" that they could build a narrative around. I just looked at the charts to see if there was any obvious discrepancies. There were.
In the chart you show, what catches my eye immediately is the totally anomalous drop of non-COVID deaths below the orange line IMMEDIATELY after the peak of COVID deaths in roughly February(?) of 2021. As a student of mortality statistical probability going back to John Graunt in the 17th century, I can assure you that didn't happen. Notice that in the earlier two peaks, the "non-COVID" excess deaths rose in tandem with the COVID deaths. This time, they inexplicably dropped right after the peak? Nope. Not buying it.
What I think you have there is progressive accumulation of reporting delays. In other words, "noise." Unless you can explain why non-COVID deaths dropped in the wake of a COVID peak, statistical noise is the most reasonable explanation.
totally read this
1. Death certificates "do not go into the data." Fifty-seven reporting jurisdictions report data to the National Center for Health Statistics based on death certificates. I doubt NCHS ever sees the death certificates. "Late filing does not create noise" is not what the NCHS statement about data limitations says. The data are weighted with estimates to correct for underreporting but those estimates cannot account for variability in cause of death.
2. It is noted that you didn't even acknowledge, let alone try to explain the SPECIFIC anomaly I pointed out, which is about the same size as the anomaly you are hawking. In my experience with statistics, cherry picking data is not productive of good analysis. I referred to above and below the orange line (Upper Bound Threshold) because that is what you showed in your chart. I am not at all surprised that Non-COVID deaths would be elevated above "Expected Deaths" because during a prolonged pandemic, many people postpone or are denied medical treatment that in normal times might prevent a future death. Because "Expected Deaths" is estimated based on non-pandemic conditions, it doesn't literally mean what would be expected NOW.
3. My "unintended" compliment was actually an intended one. I read and took seriously your deductions from the data and then considered whether there might be limitations to the data that would render your deductions as more like speculation. I would be very grateful if readers gave the same degree of consideration to my observations.
Fair enough. There is nothing wrong with spotting a possible emerging trend and calling attention to it. I don't think you have jumped to conclusions but have mooted several possible explanations, all of which are at least plausible. I would add, though, noise in the data as another plausible explanation. I wouldn't be shocked about silence from the CDC. Anything they say -- literally anything -- is subject to distortion, manipulation, and propagandizing. If they said the data is too incomplete they would be accused of concealing the facts. etc., etc., etc. This is not to say institutions do not conceal information that puts them in a bad light. I do have several epidemiologists in my immediate circle of friends who are notoriously critical of Big Pharma and consequently skeptical of the actions of the government bodies that regulate drugs. On balance, I've found the official pronouncements more credible than the alternative views on this one. Given the choice between the CDC and an "Anarcho-Libertarian" propaganda website, I would go with the former. I view right-wing libertarianism as a gateway drug to white supremacism.
BOOOOOOOO CAP lier
connect the dots...
"'What will people do if they are thrown out of work because of new technology, immigrants, or free trade?' It’s a common question that, too often, rests on the “lump of labor” fallacy; namely, the idea that there is only so much work to be done and no more, and that if one person takes a job then someone else is necessarily out of luck."
yeah fax go get em tiger