(1) do not make absolutist statements without understanding the nature of the data; (2) Do not abuse statistical terminology; (3) do not assert a conspiracy remains in location just since the information do not conform to your preferred story.
First, think about a talk about the Hurricane Maria death toll:
Excess deaths in PR through year end, those tape-recorded by the Statistics Office, numbered only 654. While the power outages there were worsened by the state ownership of PRs energy, a big portion of the excess deaths would likely have actually occurred regardless, offered the terrain and the strength of the typhoon. Hence, perhaps 300-400 of the excess deaths would have taken place regardless of steps anybody might have made to fix the power supply.
I would note that excess deaths fell by half in December. Therefore, the data suggests that the cyclone accelerated the deaths of ill and passing away people, rather than killing them outright. I would anticipate the excess deaths at a year horizon (through, say, Oct. 1, 2018) to total maybe 200-400. Still a noteworthy number, but certainly not 4,600.
See the analysis: https://www.princetonpolicy.com/ppa-blog/2018/5/30/reports-of-death-in-puerto-rico-are-wildly-exaggerated
I would note that the official death toll is 2975, in GWU report commissioned by the Commonwealth of Puerto Rico, see discussion of estimates here.
Second, a 2018 post regarding unpredictability in analytical reasoning.
Mr. Steven Kopits takes concern with the Harvard School of Public Health led studys point quote of (4645) and self-confidence period (798, 8498) for Puerto Rico excess casualties post-Maria thusly:
Does Harvard guarantee the research study, or not?
That is, does Harvard SPH believe that the central quote of excess deaths to 12/31 is 4645, or not? Is there still a 50+ probably that the death toll comes in over 4600?
Or should we simply take whatever number HSPH publishes in the future and divide by 3 to get a realistic quote of actual?
Lets show a detail of the chart formerly displayed (in this post):.
Figure 1: Estimates from Santos-Lozada and Jeffrey Howard (Nov. 2017) for September and October (computed as distinction of midpoint price quotes), and Nashant Kishore et al. (May 2018) for December 2017 (blue triangles), and Roberto Rivera and Wolfgang Rolke (Feb. 2018) (red square), and Santos-Lozada price quote based on administrative data released 6/1 (large dark blue triangle), end-of-month figures, all on log scale. Orange triangle is Steven Kopits estimate for year-end as of June 4.
The middle paragraph (highlighted red) shows a misconception of what a confidence interval is. The true specification is either in or not in the self-confidence period. Rather, this would be a much better characterization of a 95% CI:.
” Were this treatment to be duplicated on numerous samples, the fraction of computed confidence intervals (which would differ for each sample) that encompass the true population specification would tend towards 95%.”.
In other words, it is a mistake to say there must be a 50% possibility that the actual number will be above the point quote. That is exactly what Mr. Kopits believes a self-confidence period indicates.
University of Puerto Rico statistician Roberto Rivera, who along with colleague Wolfgang Rolke used death certificates to estimate a much lower death count, stated that indirect estimates ought to be analyzed with care.
” Note that according to the research study the true variety of deaths due to Maria can be any number between 793 and 8,498: 4,645 is not most likely than any other worth in the range,” Rivera stated.
As soon as once again, I believe it finest that those who want to talk about price quotes need to be familiar with analytical ideas.
Here is a reprise of a current post.
Reader Steve Kopits discusses the dispute over work numbers:.
At the same time, I thought it possible that both studies were in reality appropriate, however garbled with the effect of the healing from the suppression, therefore producing misleading impressions because we were misinterpreting the information. That still appears possible, though Ive checked out that others believe the CES was controlled to offer a more rosy image heading into the election
.
This statement joins a long pile of such accusations, e.g., Senator Barraso, Jack Welch, former Rep. Allan West, Zerohedge, Mick Mulvaney, to name a few. All I can state is that if there was a conspiracy, they didnt do an extremely great job. With the benefit of the January benchmark revision, we can update our evaluation of how severely the purported conspirators performed their job.
Excess deaths in PR through year end, those recorded by the Statistics Office, numbered only 654. While the power outages there were intensified by the state ownership of PRs energy, a big portion of the excess deaths would likely have actually happened regardless, provided the terrain and the strength of the cyclone. Therefore, the information recommends that the hurricane sped up the deaths of ill and dying individuals, rather than killing them outright. That is, does Harvard SPH believe that the main quote of excess deaths to 12/31 is 4645, or not? Figure 1: Estimates from Santos-Lozada and Jeffrey Howard (Nov. 2017) for September and October (computed as difference of midpoint quotes), and Nashant Kishore et al. (May 2018) for December 2017 (blue triangles), and Roberto Rivera and Wolfgang Rolke (Feb. 2018) (red square), and Santos-Lozada estimate based on administrative data released 6/1 (large dark blue triangle), end-of-month figures, all on log scale.
Figure 1: Nonfarm payroll employment in January 2023 release (red), in October 2022 release (blue), in 000s, s.a. Source: BLS via FRED.
Now, it might end up eventually (after another benchmark revision the results of which will be launched in February 2024) that in Q2 NFP will end up being lower than suggested in the CES. But for purposes of deceiving the electorate in November 2022, this looks like a lousy method of doing it.
In any case, before individuals begin weeping that the data are controlled, I wish they would check out the BLS technical notes on (1) revisions and indicate outright modifications, (2) benchmark modifications, (3) the estimation of seasonal adjustment elements, (4) the application of population controls in the CPS. Prior to they begin mentioning the various series, I want they comprehended the informative material (relative to business cycle fluctuations) of the CPS work series vs. that of the CES work series. That understanding can be gotten by checking out works by individuals who comprehend the qualities of the macro information (Furman (2016 ); CEA (2017 ); Goto et al. (2021 )).
From a sociological point of view, I do question why conspiracy theories are so appealing to some individuals. Heres a Scientific American post setting out some of the character characteristics that are related to adherence to conspiracy theories.
f
.
