The Data Don't Lie (But the Media Do)

Incoming: truth bomb.

If you pay attention to the news, all you hear is that COVID cases and deaths are "surging," especially in those bad red states like Florida, Texas, Georgia and South Carolina, where the evil Republican governors re-opened their states too soon, cavalierly imposing death sentences on thousands of their constituents. (Even if you ignore the news like I do, if it's on in the background for more than five minutes, you can't miss this relentless barrage.)

As always, the Curmudgeon seeks to test such hypotheses using the real data, to see whether it's true, or whether the media are lying to you again. And - guess what?

The media are lying to you again.

Yes, cases are increasing in many states now that they've opened up. Remember the objective back in March? "Flatten the curve." And flatten the curve, we did. So states began to re-open, some (including those listed above) as early as the beginning of May.

Now, cases in those states (actually nearly all states) are increasing. That makes sense. With every passing day, cases will rise. However, in some locations, they'll rise more than in others. So is what the media are screaming about true?

No, it is not. Regular readers may recall that the Curmudgeon spent the first half of April debunking the IHME COVID model, which infamously predicted 2.2 million deaths in the U.S. back in February. Because it was so clear that the model was so starved for data that it was dead in terms of any predictive value, I stopped running those analyses. In fact, I haven't updated my data since April 17.

So with the media crowing about the impending doom we all face, and the utter failure of you-know-who in battling COVID, I decided to freshen up the data as of yesterday, July 11.

And what I found was eye-opening.

Lest you mistrust the Curmudgeon, the data are presented below. I'll let you pause to peruse it before I sum up. Note that where the deaths/cases ratio is better as of July 11, that ratio is in green font. In states where the July ratio is higher than April's, it's in red. If they're equal, it's in black.

State	7/11 Cases	4/17 Cases	7/11 Deaths	4/11 Deaths	7/11 Deaths/ Cases	4/17 Deaths/ Cases
AL	51,910	4,723	1,113	147	2.1%	3.1%
AK	1,385	314	17	9	1.2%	2.9%
AZ	119,930	4,719	2,151	177	1.8%	3.8%
AR	27,864	1,777	313	38	1.1%	2.1%
CA	318,941	30,811	7,021	1,148	2.2%	2.7%
CO	36,591	9,433	1,725	411	4.7%	4.4%
CT	47,287	17,550	4,348	1,086	9.2%	6.2%
DE	12,743	2,538	517	67	4.1%	2.6%
DC	10,801	2,666	568	91	5.3%	3.4%
FL	254,511	25,492	4,197	748	1.6%	2.9%
GA	114,401	17,841	2,996	677	2.6%	3.8%
HI	1,200	574	19	9	1.6%	1.6%
ID	9,928	1,668	101	44	1.0%	2.6%
IL	154,094	29,160	7,369	1,259	4.8%	4.3%
IN	51,079	10,641	2,756	545	5.4%	5.1%
IA	34,649	2,513	748	74	2.2%	2.9%
KS	18,890	4,419	294	134	1.6%	3.0%
KY	19,121	2,707	622	144	3.3%	5.3%
LA	76,803	23,580	3,408	1,267	4.4%	5.4%
ME	3,520	847	112	32	3.2%	3.8%
MD	72,647	12,308	3,310	463	4.6%	3.8%
MA	111,398	36,372	8,310	1,560	7.5%	4.3%
MI	76,370	30,971	6,313	2,308	8.3%	7.5%
MN	41,571	2,213	1,537	121	3.7%	5.5%
MS	35,419	3,974	1,230	152	3.5%	3.8%
MO	28,475	5,517	1,117	184	3.9%	3.3%
MT	1,677	426	29	10	1.7%	2.3%
NE	20,777	1,287	286	28	1.4%	2.2%
NV	26,838	3,626	582	151	2.2%	4.2%
NH	6,024	1,342	391	38	6.5%	2.8%
NJ	180,608	81,240	15,595	4,070	8.6%	5.0%
NM	14,549	1,798	539	53	3.7%	2.9%
NY	426,798	241,041	32,388	17,671	7.6%	7.3%
NC	83,820	6,366	1,523	185	1.8%	2.9%
ND	4,243	528	87	9	2.1%	1.7%
OH	64,230	10,222	3,041	451	4.7%	4.4%
OK	19,779	2,570	421	139	2.1%	5.4%
OR	11,851	1,844	232	72	2.0%	3.9%
PA	99,229	31,731	6,953	1,102	7.0%	3.5%
RI	17,312	4,491	976	137	5.6%	3.1%
SC	54,699	4,426	951	119	1.7%	2.8%
SD	7,454	1,542	109	7	1.5%	0.5%
TN	61,006	6,762	738	145	1.2%	2.1%
TX	257,359	18,679	3,203	469	1.2%	2.5%
UT	28,855	2,931	212	25	0.7%	0.9%
VT	1,283	803	56	38	4.4%	4.7%
VA	69,782	8,053	1,962	258	2.8%	3.2%
WA	40,782	11,802	1,424	624	3.5%	5.3%
WV	4,146	825	96	18	2.3%	2.2%
WI	35,679	4,199	821	211	2.3%	5.0%
WY	1,829	422	21	2	1.1%	0.5%
U.S.	3,351,303	738,923	137,349	39,015	4.1%	5.3%

In the words of Master Po, "What have we learned today, Grasshopper?" (Google it if you're under 45).

Well, the states that have come under fire for re-opening too soon - the aforementioned red states, plus a few states with Dem governors, including California and North Carolina - have seen their deaths/cases decline since mid-April. In some cases, by half.

Where have we seen deaths/cases increase? In the states the media praise: New York, New Jersey, Pennsylvania, Michigan, Delaware, D.C., Massachusetts, Connecticut, to name several. (I know, D.C. isn't a state. If it were, it would be the State of Dysfunction.)

Now, if the Curmudgeon were as partisan as most of those who disagree with him, he'd just leave it there. But actually, there are some sound statistical reasons why those states have seen such increases - in some cases, double the April rate. States like New York and New Jersey have, tragically, seen huge numbers of deaths - far more than anywhere else in the U.S., or the world Those states' deaths were peaking around mid-April. The peaks were so high, that as the death count per day gradually began to decline, the numbers were still very large, relative to the rest of the U.S. A number of the other states didn't peak until early to mid-May. Had I been tracking the data weekly since mid-April, we'd probably see improvement in places like New York and New Jersey since maybe late May.

Of course, I'd be remiss if I didn't also mention that the governors of some of those states made major policy blunders that led to increased deaths, like releasing elderly patients from the hospital directly back to their nursing homes, where they infected many of their fellow residents. That resulted in far more deaths, as that population is the most vulnerable.

Another fallacy in the data is that entire states are treated as though the increase in cases is the same state-wide, when the sharpest increases are actually isolated. In Florida, it's largely the Miami area. In Texas, Houston is faring worse than most of the state.

I'm writing this from New Mexico, where the governor recently re-closed indoor dining, limited patio dining to 50% capacity, and required masks be worn constantly, even while exercising. She also has required that visitors from out of state - any state, even one with numbers well below New Mexico's - self-quarantine for 14 days. They can't even go out for food. She has also banned out of state visitors from New Mexico's state parks.

Are tourists really the problem in New Mexico? The state did recently see an increase in cases. However, the vast majority of those cases were in three counties: the one that includes Albuquerque, the state's largest city; the one that includes the largest portion of the Navajo Nation in the state; and the one that includes one of the state's largest prisons - and sits on the Mexico border. We know that, sadly, Native Americans and Latinos are more vulnerable to the virus. And it tends to spread rapidly among prison populations. Taos County, where I'm currently sitting, has seen 52 cases and one death since March. And Taos County is probably the most popular visitor destination in the state.

Sorry, guv. I don't think this Kansas boy is your biggest threat. If anything, visiting your state is more of a threat to me.

Now, note that several states' July deaths/cases ratios are worse (higher) than the April ratio, and thus the July number is in red font - but it's italicized. They are the Dakotas, West Virginia and Wyoming. Why the italics? If you've studied statistics, you understand the "law of large numbers." In this case, those states' data suffer from the opposite: their numbers are too small (including the state population) for the differences to be relevant. The same can be said for Hawaii, where the deaths/cases rate was unchanged from April to July.

"Aha!" the naysayers cry, "that can work both ways! A small improvement in the ratio from April to July might be just as meaningless!" Well, they can say that - but it's not true. In any state that saw an improvement, if you look at the magnitude of that improvement, however small, and compare it to the state's population, it's significant.

So what about California, Florida, Texas? How have they seen improvement in the deaths/cases ratio, in spite of increasing cases and deaths? There are a few simple reasons:

1. Generally speaking, deaths in those states are rising more slowly than cases. That's a function of the next two points.
2. More people in younger, less at-risk age cohorts are increasingly getting infected, because they tend to gather in large groups more, especially indoors. ("But they still might die!" say the naysayers. Yes - and many of them have co-morbidities that increase their risk, regardless of age. They could also die from the flu. And - the objective was never to make sure that not another human being gets infected by COVID. Is one life too many? Sure. So is one suicide, one death at the hands of a drunk driver - including the drunk driver's own life. So is one inner-city shooting. One drug overdose. Etc.)
3. Cases are increasing sharply because, in part, of increased testing. How does that work? Well, we weren't testing back in February and March (nor much in April). But there were probably a lot of asymptomatic cases that we didn't know about, because we weren't testing. Don't believe me? Believe the CDC. They recently estimated that actual infections are at least ten times confirmed cases, meaning that as much as 10% of the population may have been infected. Oh, I know, testing can't be the reason, because Pres. Trump said it was, so that has to be false. Don Lemon said so. Well, consider this: California is #1 in the nation in testing. Texas is #3. Florida is #4. Georgia is #8. North Carolina, #9. Starting to get the picture?
4. Hospitals are better prepared to deal with the virus than they were in April, so recoveries are probably higher (but we'll never know, because no one reports accurate recoveries, still). The fact that the length of hospital stays is down by half might be evidence of this.

One final consideration: the Curmudgeon has said all along, regarding data on the virus, that denominators matter. Meaning we have to look at the numbers relative to population. Now, I considered looking at changes in cases per capita and deaths per capita since April. But that would be relatively meaningless: as cases increase, cases per capita increase. As deaths increase, deaths per capita increase. And that's happening in Florida, Texas, New York and New Jersey.

However, we can look at where we are today with respect to deaths per capita (or, for a more useful measure, deaths per 1 million persons). Here's a partial breakdown (including some "hot spot" countries, for reference):

New Jersey: 1,756
New York: 1,665
Connecticut: 1,220
Massachusetts: 1,196
D.C.: 805
Louisiana: 733
U.K.: 660
Michigan: 632
Spain: 607
Italy: 578
Maryland: 548
Pennsylvania: 543
U.S. average: 420
Georgia: 282
Florida: 195
South Carolina: 185
North Carolina: 145
Texas: 111

See the differences? The "top" 5 U.S. states, including D.C., have numbers that are worse than any of the European hot spots. New Jersey and New York have numbers nearly three times those hot spots - and six to eight times those of the "bad" red states that supposedly re-opened too early.

Bottom line? You know what the media are trying to do. (Hint: it has something to do with November 3. To hedge their bets, there's always RussianbountiesRogerStonetaxreturnsracismreopeningschools.) Be smarter than that. Look at the data. Make up your own mind.

Remember that, in the end, sheep get slaughtered - by the very hand that feeds them.