Someone help me understand this

[*PhilHos*]

Look at the graphic below. I saw it on reddit this morning. I figure it's an attempt at spinning the COVID death numbers to make it seem like most of the 200,000+ deaths were inevitable.

I honestly don't know what to think. It seems convincing and while there were those on reddit trying to explain why it's misleading and/or wrong, they were too confusing.

So I'm asking, in all sincerity, what, if anything is wrong with this graphic:

[Au165]

The issue is your data seems off. I just pulled the 2019 death by month data, I then pulled the provisional death counts here (https://www.cdc.gov/nchs/nvss/vsrr/COVID19/index.htm). I then used the deaths by all causes data by month and laid them against the corresponding months from 2019. Doing this we come to an excess death count of 273k from February through September.

Now, January isn't included in the provisional death counts because it's as it related to COVID and obviously COVID wasn't really being tracked here in January and therefor they didn't include it in their chart. However, even if I gave 2020 the most favorable of outcomes of 30k better than 2019 (2x 2020's best month) you'd still be around 240k deaths ahead of 2019.

The idiot who made your chart probably used the running provisional death count total (based on where it was on October 3rd), the issue was he didn't recognize that January was missing from that total, while all his other stats included January. When looking at the 3 year average of January missing at 269k people you then magically realize...that is where the 200k people went.

[fredtoast]

Have to see the numbers broken down more to decide.

Maybe 200K fewer people died from heart disease, accidents, cancer, or other things that could not be attributed to Covid in 2020.

Or maybe the numbers just are not correct. I have seen a lot of complete bullshit lies posted as fact on Reddit before.

I just can't tell from these stats.

I do know that I have seen a lot of analysis of "excessive deaths" that show more deaths form categories that could include Covid deaths.

[BigPapaKain]

If you beat the numbers long enough, they'll say whatever you want them to say.

Especially when you use false data.

[Au165]

(10-08-2020, 01:55 PM)fredtoast Wrote: Have to see the numbers broken down more to decide.

Maybe 200K fewer people died from heart disease, accidents, cancer, or other things that could not be attributed to Covid in 2020.

Or maybe the numbers just are not correct.  I have seen a lot of complete bullshit lies posted as fact on Reddit before.

I just can't tell from these stats.

I do know that I have seen a lot of analysis of "excessive deaths" that show more deaths form categories that could include Covid deaths.

#3 they unintentionally, or potentially intentionally, left out January in this year's data which accounts for roughly 269k people on average each year.

[*PhilHos*]

(10-08-2020, 01:55 PM)Au165 Wrote: The issue is your data seems off. I just pulled the 2019 death by month data, I then pulled the provisional death counts here (https://www.cdc.gov/nchs/nvss/vsrr/COVID19/index.htm). I then used the deaths by all causes data by month and laid them against the corresponding months from 2019. Doing this we come to an excess death count of 273k from February through September.

Now, January isn't included in the provisional death counts because it's as it related to COVID and obviously COVID wasn't really being tracked here in January and therefor they didn't include it in their chart. However, even if I gave 2020 the most favorable of outcomes of 30k better than 2019 (2x 2020's best month) you'd still be around 240k deaths ahead of 2019.

The idiot who made your chart probably used the running provisional death count total (based on where it was on October 3rd), the issue was he didn't recognize that January was missing from that total, while all his other stats included January. When looking at the 3 year average of January missing at 269k people you then magically realize...that is where the 200k people went.

See, that makes sense. Thanks. I figured there had to be something wrong because there's no way we get to 200,000+ deaths and it be the wrong number AND no one figures that out. 

(10-08-2020, 01:55 PM)fredtoast Wrote: Maybe 200K fewer people died from heart disease, accidents, cancer, or other things that could not be attributed to Covid in 2020.

That was my first thought but I would think it should still would be represented in the numbers (assuming, of course, that the numbers weren't wrong which it seems they are).

[Au165]

(10-08-2020, 02:46 PM)PhilHos Wrote: See, that makes sense. Thanks. I figured there had to be something wrong because there's no way we get to 200,000+ deaths and it be the wrong number AND no one figures that out. 

Yea, part of me thinks it was accidental because the provisional death stats start at 2/1 which probably is odd to most. The only reason I wonder if it was intentional is the way they stack that fourth column (average of last 3 years) over this year's total deaths instead of showing it month by month. Had the guy went to show the month by month breakdown for this year he would have easily seen he was missing a month.

[Belsnickel]

So, what's interesting is that I saw a report somewhere that the pandemic is also causing an increase in non-COVID deaths because people are avoiding hospitals and doctors' offices, so they aren't seeking treatment for conditions that wouldn't be fatal with treatment, but are without.

Anyway, glad we could figure out the data weirdness. I love data visualizations, and here is one straight from the CDC that actually comes from the second source link in the OP.


This shows the excess deaths over what was estimated by the CDC for the year. That is what COVID is responsible for.

[hollodero]

(10-08-2020, 03:04 PM)Belsnickel Wrote: So, what's interesting is that I saw a report somewhere that the pandemic is also causing an increase in non-COVID deaths because people are avoiding hospitals and doctors' offices, so they aren't seeking treatment for conditions that wouldn't be fatal with treatment, but are without.

Heard the same thing in my country.

Of coruse, many americans also simply lost their healthcare during the economic crisis that came with COVID.... heard there's an estimate of 5 million people affected by that. Hard to believe that doesn't come with quite a few additional fatalities.

[Benton]

Stats can be interpreted different ways, but just raw numbers don't really shed a light on much.

Maybe, overall, deaths are down because not as many people are driving as much (a big chunk of deaths, somewhere around third or fourth behind heart disease and cancer). Maybe people working from home resulted in fewer stress-related heart attacks (heart disease being the biggest killer year to year). Maybe, as some GOP candidates have alleged, nobody has really died from COVID and it's all a plot by the Trump administration to make the Trump administration look bad. Who knows?

Numbers are great, but without explanation, they're just numbers.