Predictions With and Without Confidence Intervals: Puerto Rico Post-Maria

Compare this assessment (5/31/2018):

Thus, the data suggests that the hurricane accelerated the deaths of ill and dying people, rather than killing them outright. I would expect the excess deaths at a year horizon (through, say, Oct. 1, 2018) to total perhaps 200-400. Still a notable number, but certainly not 4,600.

With a contemporaneous prediction:

From the survey data, we estimated a mortality rate of 14.3 deaths (95% confidence interval [CI], 9.8 to 18.9) per 1000 persons from September 20 through December 31, 2017. This rate yielded a total of 4645 excess deaths during this period (95% CI, 793 to 8498), equivalent to a 62% increase in the mortality rate as compared with the same period in 2016.

Subsequent to this report, the government released additional mortality data, as discussed here. A plot of estimates from that post is reproduced below:


Figure 1: Estimates from Santos-Lozada and Jeffrey Howard (Nov. 2017) for September and October (calculated as difference of midpoint estimates), and Nashant Kishore et al. (May 2018) for December 2017 (blue triangles), and Roberto Rivera and Wolfgang Rolke (Feb. 2018) (red square), and calculated excess deaths using average deaths for 2015 and 2016 compared to 2017 and 2018 using administrative data released 6/1 (black square), and Santos-Lozada estimate based on administrative data released 6/1 (large dark blue triangle), end-of-month figures, all on log scale. + indicate upper and lower bounds for 95% confidence intervals. Orange + denotes Steven Kopits 5/31 estimate for range of excess deaths through September 2018; Orange triangle is Steven Kopits estimate for year-end as of June 4. Cumulative figure for Santos-Lozada and Howard October figure author’s calculations based on reported monthly figures.

From NYT today:

Puerto Rico is now estimating that Hurricane Maria killed more than 1,400 people, far more than the official death toll of 64, in a report to Congress seeking billions to help the island recover from the devastating storm.

The government, relying on updated statistics it first reported in June, said there were 1,427 more deaths from September to December 2017 than the average for the same time period over the previous four years.

These new numbers are apparently (my inference) prompted by this JAMA letter by Santos-Lozada and Howard:

…based on death records following Hurricane Maria, the estimated hurricane-related mortality burden of 1139 excess deaths through December 2017 was higher than the official death toll of 64. The estimate is conservative, because the expected number of deaths used the upper 95% CI and did not consider the population denominators, which were decreasing. The strength of the present approach is that it is based on death counts from vital statistics records and is consistent with previous estimates1 and methods.4,5 The primary limitation of the study is that the specific cause of each individual death is not known; thus only an aggregate number of deaths in excess of historical variation can be estimated. Another recent study2 suggested that there were 4645 excess deaths (95% CI, 783-8498), but it was based on a survey that underestimated prehurricane mortality, overestimated posthurricane mortality, and had a large CI, indicating a high level of uncertainty. Future studies would benefit from careful analysis of deaths from vital records rather than surveys. [emphasis added — MDC]

From the letter:

The Puerto Rico government estimates apparently use the mean of past years rather than upper bound for the confidence interval to do the comparison. It is important to note that since the estimate Santos-Lozada and Howard estimate does not take into account population drop over the recent years (which apparently has been substantial), the average expected mortality rates are probably overstated, implying the gap between observed and counterfactual is under-stated.

While the numbers of fatalities for September-December are likely fairly firm, I don’t really know that for certain. That’s because there is a large backlog of bodies at the coroners office, predating the hurricane, prompting a request for additional mortuary officers. Interestingly, the request for additional mortuary officers came only on July 23, 2018. So, asserting that the given numbers are firm is foolhardy. (Listen also to this item).

So, there still seems to be substantial opacity regarding the gathering and reporting of mortality statistics. In these situations, it only seems right and proper to include confidence intervals. (I would also not be surprised to see the government’s estimate of excess mortality rise over time, given the incentives that governments face.)

62 thoughts on “Predictions With and Without Confidence Intervals: Puerto Rico Post-Maria

  1. pgl

    More evidence that Donald Trump and PeakRacist have lied to us:

    https://www.eurekalert.org/pub_releases/2018-08/pfan-iul080718.php

    Immigrants use little health care, subsidize care of non-immigrants: Harvard/Tufts study

    New study finds that immigrants pay more money into both private and public health insurance programs than they take out

    A study published today in the International Journal of Health Services finds that immigrants use far less health care than non-immigrants, and may actually subsidize the care of U.S. citizens. The findings by researchers at Harvard Medical School and Tufts University School of Medicine contradict recent claims by Pres. Trump and other political leaders that immigrants drive up U.S. medical costs.
    Researchers examined all studies published since 2000 related to health care expenditures by immigrants. Compared to U.S.-born individuals, immigrants were found to have lower rates of utilization and lower per capita expenditures from private and public insurance sources; health expenditures were particularly low for undocumented immigrants. Across all age groups, immigrants’ utilization was only one-half to two-thirds as high as that of the U.S.-born population. Immigrants also made larger out-of-pocket health care payments compared to those born in the U.S.

    Reply
  2. baffling

    so now the death toll begins to approach the numbers we saw from hurricane katrina. yet there were those on this blog who refused to acknowledge the comparison, and what should be at least similar order of magnitude death tolls. there was an order of magnitude difference between the two death tolls, but that did not seem to concern some of the defenders of the government numbers. unfortunately, i ultimately expect the maria death toll to be higher than katrina.

    Reply
    1. pgl

      I bet Princeton Steve’s “research” would be less government friendly if this disaster had occurred under President Obama.

      Reply
      1. Steven Kopits

        The numbers are what they are. Survey data , except in places where governance has completely collapsed (eg, Syria), will always lose out to actual death certificates

        There is, however, a case to be argued that Hurricane Maria led to 4600 pre-mature deaths, but that’s not the same as excess deaths.

        Reply
        1. Menzie Chinn Post author

          Steven Kopits: One of the first things you learn in government (in my case CEA) and quasi-government (IMF) is … don’t accept numbers at face value. Even a year after Maria landfall, it’s not clear to me that the government has tabulated all the dead. And, remember, the Santos-Lozada/Howard estimates did not account for the depletion of population of the year immediately before Maria, as the economy was declining rapidly.

          Reply
          1. Steven Kopits

            Menzie,

            These are dead people, not more abstract quantities like trade balances of GDP.

            Dead people have certain characteristics. For example. people need to bury them. That means the ME has to release the body for burial. It means a funeral home has to process the body. All this takes paperwork, and it has to be completed pretty quickly. Similarly, if people are missing, they are missed and the reports about the lists of the missing will be all over the press. And they’ll know where the missing are buried (eg, in a landslide) and be digging for them. All this leaves a trail in the bureaucracy or in the press.

            If you think the issue is backed up paperwork, and you’re the Harvard team, you pick up the effing phone and call a few ME’s and ask them what’s going on. That’s exactly what I mean by saying there was no senior management oversight on this project. It’s the project manager’s job to read the damn report and say, “Well, as of Jan. 2, we have about 1000 bodies we can account for. Where are the other 3,600? Until someone can show me where these bodies could plausibly be, no way in hell we publish a report which says 4.645 excess people could have died as a result of the storm.” That step never happened, and that’s what make the whole study garbage.

          2. Menzie Chinn Post author

            Steven Kopits: Thanks. Your analysis is my case study for next time they let me teach Public Affairs 819, statistical analysis for public affairs.

          3. Steven Kopits

            That’s what’s wrong with statistical analysis. It’s not your only source of data! If you don’t sense check your analysis, you may produce gibberish and prompt a rebuke from Jama along the lines of “Future studies would benefit from careful analysis of deaths from vital records rather than surveys.”

          4. Menzie Chinn Post author

            Steven Kopits: Let’s get this right. It’s not a rebuke from JAMA. It’s a rebuke from another academic, who is submitting a letter to JAMA. A lot like the things I write about other people’s papers, but I do not in my arrogance and hubris call garbage.

          5. Moses Herzog

            Verbatim from “Princeton” Kopits: “Dead people have certain characteristics. For example. people need to bury them. That means the ME has to release the body for burial. It means a funeral home has to process the body. All this takes paperwork, and it has to be completed pretty quickly.”

            Curious how “Princeton”Kopits thinks people make those “pretty quick” phone calls when communication/phone networks are down?? How do they find those bodies in areas where roads and bridges have been destroyed?? How do they find those bodies when they are lying underneath destroyed structures?? Do Puerto Ricans achieve these feats with “magical” paper towels that trump tossed at them in a crass photo-op designed for people with “Princeton”Kopits’ low level of intellect??

            In “Princeton”Kopits magical world of the documentation and counting of dead bodies, when someone hasn’t been found or even accounted for as missing, how do you get an accurate count?? That’s “easy” to solve in “Princeton” Kopits’ world. Some braindead Kopits type just yells out loudly “the ME has to release the body for burial!!!! And also, by the way, the funeral home has to process the body!!!! FYI from the white guy!!!!” and the problem (with the assistance of trump’s “magical” paper towels) is solved.

          6. Steven Kopits

            Moses,

            The report was issued on May 29, eight months after the hurricane. The phones were up in PR by then.

            Further, there was a big team from the Milken Institute of GW literally combing through the death certificates at the time. The Harvard team could simply have called them up. Further, the Harvard team did not call the Governor’s office, which it also could have done. Major, major errors of project management.

            https://publichealth.gwu.edu/content/milken-institute-sph-project-will-estimate-excess-mortality-puerto-rico-tied-hurricane-maria

          7. pgl

            “Dead people have certain characteristics. For example. people need to bury them. That means the ME has to release the body for burial. It means a funeral home has to process the body. All this takes paperwork, and it has to be completed pretty quickly.”

            You know – these little lectures from Princeton Steve have gotten absurd. The reality on the ground is that there are dead people in Puerto Rico (US citizens) who are being stored in refrigerated trucks as the system of identifying them and properly burying them are overwhelmed. And yet Princeton Steve seems to be totally unaware of the reality on the ground.

            And yet this pompous fool thinks his “research” is definitive while that of others is “garbage”.

          8. baffling

            “Further, there was a big team from the Milken Institute of GW literally combing through the death certificates at the time. The Harvard team could simply have called them up. Further, the Harvard team did not call the Governor’s office, which it also could have done. Major, major errors of project management.”

            first, steven, you think the milken team would have given up their number to another “competing” research group prior to publication? how naive are you? hey intel, i am working on a new, faster cpu that will change the world. but i am not sure about a few of the details of how it works. since you are working on something similar, do you mind if a take a look at your work and compare it to mine, for safety’s sake? second, the PR government has been adamant about their official number for months. you think that phone call would have resulted in a different response?

            steven, i get the picture you are promoting for academic research to have professional managers run the show-somebody like yourself? i fail to see how that improves the picture in any way whatsoever. final approval and publication is dependent upon a “professional” with no expertise in the area and operates off of a “hunch”. no thanks. real research does not work that way.

        2. baffling

          “There is, however, a case to be argued that Hurricane Maria led to 4600 pre-mature deaths, but that’s not the same as excess deaths.”
          i fail to see the distinction here steven. if i was one of the excess deaths caused by hurricane maria, i would certainly consider my death pre-mature. you either die as a result of the storm, or you do not. takes a lot of hubris to state somebody who died from the storm is not really a victim because, in YOUR opinion, he would have died soon anyways. i suppose we now have a pre-exiting illness clause in the death count machinery.

          Reply
    2. CoRev

      Baffled, will you list the definitive count and its reference for each event, please. IIRC, even the Katrina count after several years was an estimate in which interest had waned to pursue further.

      Reply
        1. CoRev

          Baffled, the point was: ”
          baffling
          August 9, 2018 at 11:34 am

          so now the death toll begins to approach the numbers we saw from hurricane katrina. yet there were those on this blog who refused to acknowledge the comparison, and what should be at least similar order of magnitude death tolls. there was an order of magnitude difference between the two death tolls, but that did not seem to concern some of the defenders of the government numbers. unfortunately, i ultimately expect the maria death toll to be higher than katrina.”
          Compare the two event counts.

          Reply
          1. baffling

            they should be similar. i would imagine maria would be the worse of the two. currently, they do appear to be similar. this is why using the value of 64 in any discussion should have raised a red flag. this comparison was obvious in my previous post, unless you are illiterate.

          2. CoRev

            Baffled, so you are BSing again without knowing the actual counts. And you call others names???

          3. baffling

            corev, you are simply baiting. you will complain about any number i provide. but again, not even sure what you are complaining about, other than to simply be the party of no. if i were to say the grass is green, you would have to disagree with me out of principle. idiot.

          4. CoRev

            Baffled, baiting??? Me???? Why would I ever do that, to you about something so serious as deaths from hurricanes? How does it feel?

            I remember a past thread where you tried the same approach and even you references for Katrina deaths admitted they were incomplete estimates, and the effort to clarify had been stopped after 3 years. All of that went on while we were discussing the inability to get better estimates in the early months after Maria.

          5. baffling

            corev you are baiting because there would never be an acceptable answer in your opinion. you would rather have no estimates than anything that would question your world view. we all know death counts are estimates. again, i really don’t understand what you are complaining about, other than to be the party of no. i state that the grass is green, what is your response coloser? you are nothing but a lonely idiot.

  3. Steven Kopits

    “Another recent study2 suggested that there were 4645 excess deaths (95% CI, 783-8498), but it was based on a survey that underestimated prehurricane mortality, overestimated posthurricane mortality, and had a large CI, indicating a high level of uncertainty. Future studies would benefit from careful analysis of deaths from vital records rather than surveys.”

    https://jamanetwork.com/journals/jama/fullarticle/2696479

    In other words, the conclusions of the Harvard study were garbage. To be clear, nothing wrong with trying a survey approach, but as Jama notes, “Future studies would benefit from careful analysis of deaths from vital records rather than surveys.”

    Reply
    1. Menzie Chinn Post author

      Steven Kopits: You 5/31 “interval” of 200-400 did not encompass the current 1400 estimate. The Harvard SPH study did. And you say theirs is garbage?

      Reply
      1. Steven Kopits

        Well, we don’t know whether mine will be right because it is at the year horizon, Oct. 1, as I recall.

        But you’ll also recall, because this is the fifth time at least that I have written it, that the PR official data as of Jan. 2 were ‘as of’, which I believed at the time to be ‘preliminary’. Mea culpa. If you like, I’ll copy paste this again to make it a round half dozen. Or will five do?

        The 1400 number proved a rock solid estimate, as we knew it would. https://www.nytimes.com/2018/08/09/us/puerto-rico-death-toll-maria.html

        The Harvard results were garbage, and the way the findings were presented was even more garbage. There was no senior project management oversight on the MPR report, despite 15 authors of which 13 had advanced degrees (or more realistically, because there were 13 authors with advanced degrees). Or they let some anti-Trump ideology get the better of them.

        Reply
        1. Menzie Chinn Post author

          Steve Kopits: But that’s part of the problem. When one knows the data is likely to be contaminated or biased, and there is information that suggests incentives to downplay numbers, then one needs to (1) be wary of the reported data, be they official or not, and (2) use alternative approaches to infer the relevant range. You did neither. You took as given the official numbers, and ran with them.

          A post-facto mea culpa is not sufficient. What you need to do is say one should (1) acknowledge uncertainty regarding reporting errors, (2) acknowledge statistical uncertainty, and (3) not describe as “garbage” a study that uses an appropriate methodology, and has the honesty to report the degree of uncertainty, in debating the validity of estimates.

          By the way, I don’t know what the citation of October 1st window is supposed to prompt. Do you actually think the estimates of excess deaths will go down with additional data, as the additional staffing for the coroner’s office comes in, and additional bodies are found over time?

          Reply
          1. Steven Kopits

            Menzie, I am not an expert in Puerto Rican mortality, but for god’s sake, you could pick right off the page. I was unable to make a judgment about the quality of the PR data because I had only one vintage (or was it two?) — that in the study itself. I was unable to determine what those numbers meant (ie, ‘as of’ v ‘preliminary’ v ‘final’) until the CBS guy released the updated numbers a few days later. But it was readily apparent that it was highly, highly unlikely that the excess death toll would come anywhere near 4,645. And it didn’t!

            Also, keep in mind that excess deaths are as of a date. We believe that the hurricane brought forward the deaths of a vulnerable population dependent on air conditioning, respirators, dialysis and other life support systems. These people were not in good shape to begin with, and therefore we would expect that some portion of them would have died relatively soon in any event. Therefore, the excess death number should peak and then decline as the time frame is extended. In the year 2200, the number of excess deaths should be exactly zero, for example.

          2. CoRev

            Menzie a simple question. Which study do you believe is the more accurate (with confidence levels),
            Harvard? ______ or the PR Govt ____?

          3. Menzie Chinn Post author

            CoRev: No, in academia, and intelligent discussion, one usually sets the parameters of the question clearly. Please elaborate. As I currently construe your query, it makes no sense.

          4. CoRev

            MMenzie, now I see you are in bob and weave mode. I am truly surprised you didn’t just run away from a request as does pgl.

            And you obviously had no problem doing an assessment without confidence level of the PR Govt estimate: “So, there still seems to be substantial opacity regarding the gathering and reporting of mortality statistics. In these situations, it only seems right and proper to include confidence intervals. (I would also not be surprised to see the government’s estimate of excess mortality rise over time, given the incentives that governments face.)”

            As you said: “As I currently construe your query, it makes no sense.”

          5. Menzie Chinn Post author

            CoRev: Let’s be clear. The government did not provide a confidence interval as far as I can tell. It provided an estimate. I did not “assess” the government’s estimate. I did comment — I think neutrally on — the Letter to JAMA from Santos-Lozada/Howard. I understand the methodology they pursued. I understand the methodology the team let by Harvard SPH researchers undertook. When the Harvard SPH study came out, they did not have the extended mortality data that is contained in the Santos-Lozada/Howard study. Rather, they had the same mortality data that Steven Kopits used to make his no more than 200-400 deaths by October 1, 2018 estimate.

            There are (statistical) biases one can see in both studies; these biases are very hard to correct (I’m trying now, and it’s hard). In the Harvard SPH study, one also gets a very large confidence interval. Based on the data we now have access to, I would probably put more weight on the Santos-Lozada/Howard tabulation. However, given the recent revelations regarding the understaffing of the Coroners Office in Puerto Rico, you too might want to downweight official mortality data.

            (By the way, all those people who worried about Federal government manipulating official data — why are they now so keen to believe w/o question government data?)

            So, given this explanation, let me re-iterate: your question as posed makes no sense because as far as I have seen the government of Puerto Rico has not released a confidence interval associated with its revised estimate.

          6. baffling

            menzie, as i recall, coloser was a real stickler for confidence intervals and error bands when he argued that yearly temperatures were not record setting. now he seems to have lost interest in those items. ironic to say the least.

          7. Menzie Chinn Post author

            baffling: Yes, it is interesting that CoRev is focused almost primarily on economic issues these days. I think the confidence interval issue is “NLC” — or “no longer convenient” for him/her.

          8. ottnott

            Kopits wrote:
            Also, keep in mind that excess deaths are as of a date. We believe that the hurricane brought forward the deaths of a vulnerable population dependent on air conditioning, respirators, dialysis and other life support systems. These people were not in good shape to begin with, and therefore we would expect that some portion of them would have died relatively soon in any event. Therefore, the excess death number should peak and then decline as the time frame is extended. In the year 2200, the number of excess deaths should be exactly zero, for example.

            Stunning.

            All that blather about how to properly carry out a study and you end with a comment totally dismissive of the reasons we bother trying to figure out a storm’s damage in terms of human deaths.

          9. Menzie Chinn Post author

            ottnot: I agree it’s pretty stunning. I have been thinking about this point, as it does (as was noted) imply excess deaths from Hitler will be zero eventually. Taking the problem seriously, I suppose the alternative is to estimate a life-years loss instead of excess deaths.

        2. baffling

          “The Harvard results were garbage, and the way the findings were presented was even more garbage. There was no senior project management oversight on the MPR report, despite 15 authors of which 13 had advanced degrees”

          actually it was not garbage steven, simply producing results you did not agree with. they produced a sound scientific study which is reproducible, clearly stated their methodology, provided data and did not hide their conclusions. you simply do not like the spread in the data-but that is statistics-nothing nefarious at all. your senior project management comments are simply self serving. if a study had produced a result of 1400 deaths a year ago, and you had served as senior project manager, would you have allowed the study results to be published in light of the official count of 64? i doubt it.

          Reply
          1. Menzie Chinn Post author

            baffling: I think you’ve put your finger on the distinction between “consulting” and “academic research”. Mr. Kopits is conducting the former activity.

          2. baffling

            menzie, that has been my point with steven for several posts. i have done both, so i know it when i see it. steven needs to appreciate that academic research and professional consulting have different goals and desirable outcomes.

          3. CoRev

            Baffled says: “steven needs to appreciate that academic research and professional consulting have different goals and desirable outcomes.” Agreed! One is to further knowledge and errors are assumed, and the other is to expand or clarify a manager’s options, and errors are seldom tolerated. Which is the more valuable depends on who is paying!

          4. baffling

            “the other is to expand or clarify a manager’s options, and errors are seldom tolerated. ”
            i don’t see steven’s approach reducing the error to a more tolerable level. my guess is steven would not have approved of a report that actually captures the real numbers.

          5. CoRev

            Baffled, why the need to put words/interpret what Steven said? He has been clear in his comments, you, on the other hand, much less clear. Which number, if any, is closest to real? Please define how you determined what is real.

  4. joseph

    Kopits: “There is, however, a case to be argued that Hurricane Maria led to 4600 pre-mature deaths, but that’s not the same as excess deaths.”

    Let’s face it. Kopits is a monster. By his reasoning, 6 million Jews dying weren’t excess deaths. They were just a little pre-mature. After all, the weakest died first in the labor camps.

    Kopits: ” keep in mind that excess deaths are as of a date. We believe that the hurricane brought forward the deaths of a vulnerable population … we don’t know whether mine will be right because it is at the year horizon, Oct. 1, as I recall.”

    And Kopits is still clinging to the monstrous idea that by killing off the weak people first, you will get a deficiency of dead people in later months which he gets to subtract from the dead people killed off earlier. He claims the absolute death toll declines from its present level as these mere “pre-mature” deaths are averaged away in the future. Heck, if you go far enough in the future, there were no excess deaths at all — just “pre-mature” ones. It is a very sick mind that reasons this way.

    Reply
    1. pgl

      “I am not an expert in Puerto Rican mortality”. The first thing Princeton Steve has written that makes sense! And yet he babbles on with his incoherence!

      Reply
  5. Moses Herzog

    “The pissing match doesn’t work”. Anyone else think it’s odd hearing Paul Ryan quoting Neville Chamberlain at this point??
    https://www.nytimes.com/2018/08/07/magazine/paul-ryan-speakership-end-trump.html

    But hey kids!!!! Have I got some exciting news for you!!!! I found some illegal immigrants that PeakIgnorance, CoRev, and “Princeton”Kopits could just kiss and hug and smear their love all over them!!!!:
    https://www.washingtonpost.com/politics/melania-trumps-parents-become-naturalized-us-citizens-amid-the-presidents-hostility-toward-chain-migration/2018/08/09/646a4f62-9bf9-11e8-b60b-1c897f17e185_story.html

    Now, some of you may wonder why some immigrants make “Princeton” Kopits very angry and upset and other immigrants make “Princeton” Kopits want to break out singing Lee Greenwood’s “God Bless the USA”. Well never mind, one of “Princeton” Kopits’ in-laws is Korean. So that’s why he can’t be consistent on that stance. You understand that right?? Kopits’ in-laws are Korean, so he can be a hypocrite as long as one of his in-laws is Korean. That’s just self-apparent logic. You get it, right??

    Reply
  6. Moses Herzog

    This is NYT article is very worthy of a read through, and reading ALL of it
    https://www.nytimes.com/interactive/2018/08/09/upshot/trump-voters-how-theyve-changed.html

    Just click on the small footnote on the bottom that says you don’t wanna play the guessing game (unless you have tons of time)

    The part to me that is interesting (that NYT and/or any other major news outlets refuses to discuss) is the high number of rural women and non-college women who voted for Trump. Now, unless you want to argue a significant portion of these women are sexist towards their own gender, it really kind of kills the whole Russian conspiracy as a decisive factor in the elections. Now, I wanna make it clear here—there’s no doubt there were crimes committed by Trump campaign officials related to the Russians and the Kremlin—-but one can believe that severe and traitorous crimes were committed and tampering/collusion occurred and simultaneously believe it didn’t change the end result of the elections—which is what I myself believe.

    What honestly kind of grates on me, is how Hillary keeps trying to sell this “bill of goods” that:
    1) her knowing and intentional unlawful use of a server
    2) her psychosis of being a pathological liar
    3) her failing to campaign in Michigan, Wisconsin etc,
    4) just being a broadly unlikable candidate

    Was not the REAL cause of her losing the race.

    Reply
  7. Moses Herzog

    The difference is now 91 votes out of roughly 311,000 cast. Feel free to figure out what that equates to as a percentage:
    https://slate.com/news-and-politics/2018/08/kansas-secretary-of-state-kobach-posts-mistaken-gop-primary-vote-tally-but-somehow-still-hasnt-recused-himself.html

    No doubt still more than the number of illegal votes Kobach found in his “voter fraud task force” as the idiot quit his own investigation due to a lack of conviction in his own lies.

    Reply
  8. joseph

    Let’s not beat around the bush here. Kopits wants you to take seriously his pseudo-intellectual arguments about the deaths and destruction in Puerto Rico. “It’s just a discussion about statistics!” But we all know that this isn’t a good faith argument.

    This is about Steven Kopits providing cover for his Dear Leader’s racist depraved indifference to the lives of American citizens who don’t happen to be white. He wants to argue that liberals are over-reacting, that the deaths in Puerto Rico are exaggerated, it wasn’t really that bad.

    This is a form of trolling called “seal-lioning”. It means someone who engages in a debate with fake civility while obnoxiously and continually demanding more evidence. It’s a waste of time to engage them because there never will be enough evidence. Their objective is just want to annoy and waste your time.

    In this case it’s about defending Trump’s racist behavior.

    https://en.wikipedia.org/wiki/Sealioning

    Reply
    1. 2slugbaits

      joseph Maybe I’m just naïve, but I’ve never gotten the sense that Steven Kopits fits that description. And I don’t think he’s a Trump supporter. I believe the problem is that Steven Kopits doesn’t entirely understand or appreciate what a confidence interval actually means. Part of the problem is that floating in the ether there’s this very common ersatz freshman stat class version of how to interpret confidence intervals, but it’s really a misinterpretation.

      Reply
      1. pgl

        “Part of the problem is that floating in the ether there’s this very common ersatz freshman stat class version of how to interpret confidence intervals”

        Princeton Steve does have a rather naive understanding of even basic statistics indeed but his knowledge of statistics clearly exceeds his knowledge of economics. Which is why CoRev likes this guy!

        Reply
      2. Moses Herzog

        I tend to take Joseph’s side on this one. But I’ve been told I’m the cynical/pessimist type, so take it for what it’s worth.

        If people are consistent in their logic, then I can buy it’s an upright or well-intentioned argument. But when they apply different rules based on the parties involved, then just “being ignorant” doesn’t cut it.

        Reply
  9. Anonymous

    “I believe the problem is that Steven Kopits doesn’t entirely understand or appreciate what a confidence interval actually means.”

    Oh, I agree that Kopits is clueless about confidence intervals. But his radical argument goes far beyond that ignorance. He is saying the pre-mature deaths are different from excess deaths. And that pre-mature deaths today are cancelled out by a deficiency in deaths later. After all, nobody ever dies twice. That’s a monstrous argument.

    You can argue whether he is a Trump defender or not, but you can’t dispute that his motivation in downgrading the number of excess deaths is to minimize the claims of liberals that what happened in Puerto Rico is a tragedy.

    Reply
    1. 2slugbaits

      He is saying the pre-mature deaths are different from excess deaths.

      That’s an incoherent argument, but I don’t see that as an inherently racist argument.

      you can’t dispute that his motivation in downgrading the number of excess deaths is to minimize the claims of liberals that what happened in Puerto Rico is a tragedy.

      I have no doubt that racism is what motivates Trump and Laura Ingraham and Sean Hannity and PeakTrader in wanting to minimize the PR disaster. It’s also a refusal to admit that Trump is an incompetent bungler. I’m less convinced that racism is what motivates Steven Kopits’ estimate of the number of deaths. Instead, I think he just found it hard to believe because the Harvard numbers were so far above the official PR number. It’s a lack of faith in nerdy statistical methods when those estimates don’t agree with one’s gut feel. Every project manager I’ve ever known has always tended to overstate confidence, so the temptation is to align estimates with the client’s gut estimate. Real world confidence intervals are not always marketable. I don’t think Steven Kopits is a racist, but I do think he’s sometimes guilty of wanting to express the kind of confident conservative caution that clients like to hear.

      Reply
  10. pgl

    “i have done both, so i know it when i see it. steven needs to appreciate that academic research and professional consulting have different goals and desirable outcomes.” – baffling agreeing with Menzie.

    Look – I get that some consultants are paid to provide “evidence” that supports their clients political/partisan BS. But there are two types of professional consulting. The type alluded to here is in essence sophisticated prostitutes.

    Some clients hire professional consultants to basically tell them the Uncomfortable Truth (hat tip to Al Gore). I would argue serving as an adviser to a President can elicit this divide. Most members of the CEA were the latter type of consultants especially when the President wanted to hear the truth – even if the President did not always follow the CEA advise.

    Ah but we know the current President could care less about the real world – he just wants his agenda to be praised as the greatest policy ever. So yea – Trump wants prostitutes in his consultants. Princeton Steve would fit very well in this White House.

    Reply
  11. PeakTrader

    2slugbaits says: “I have no doubt that racism is what motivates Trump and Laura Ingraham and Sean Hannity and PeakTrader in wanting to minimize the PR disaster.”

    What incredibly poor judgement.

    It’s amazing what you get from the extremely biased lunatic fringe.

    Lazy ignorance.

    Reply

Leave a Reply

Your email address will not be published.