Debate Watch Debate Watch

Previous Month | RSS/XML | Current | Next Month


August 31st, 2020 (Permalink)

The Debate About the Debates

Will there be presidential and vice presidential debates this year? Should there be? According to Speaker of the House Nancy Pelosi, there shouldn't1. Some other Democrats and even some anti-Trump Republicans have also suggested that Democratic candidate Joe Biden shouldn't debate President Donald Trump2.

Debates for president and vice-presidential candidates have occurred in every election year since 19763. Occasionally, there has been some resistance to the debates, but usually from the incumbent rather than the challenger. This year, the reluctance to debate is coming from the challenger's supporters.

Though the Biden supporters who have called for no debates haven't mentioned it, some Republicans have suggested that the real reason for such calls is Biden's age. If he wins the election, Biden will be the oldest man ever to be elected president and, in fact, will be as old at the beginning of his term as Ronald Reagan was at the end of his two terms.

During Reagan's 1984 re-election campaign, concerns about his age were raised by his poor performance in the first debate with Democratic challenger Walter Mondale4. However, by the second debate he had recovered his form and helped to defuse the issue with a joke:

Henry Trewhitt: Mr. President, I want to raise an issue that I think has been lurking out there for 2 or 3 weeks and cast it specifically in national security terms. You already are the oldest President in history. And some of your staff say you were tired after your most recent encounter with Mr. Mondale. I recall yet that President Kennedy had to go for days on end with very little sleep during the Cuban missile crisis. Is there any doubt in your mind that you would be able to function in such circumstances?

Reagan: Not at all, Mr. Trewhitt, and I want you to know that also I will not make age an issue of this campaign. I am not going to exploit, for political purposes, my opponent's youth and inexperience.5

While the joke itself might have been the product of his speechwriters, Reagan's performance throughout the debate reassured voters enough that he was re-elected. Though President Trump is now the same age that Reagan was in 1984, the current situation differs in that Trump's advanced age is over-shadowed by that of his opponent, who is three years older, whereas Mondale was only 56 at the time.

Both campaigns have already agreed to a schedule of three presidential debates and a single vice presidential one between now and voting day. As far as I know, the Trump campaign and its supporters are not suggesting the debates be cancelled. So, if the Biden campaign followed Pelosi's advice, it would have to unilaterally back out of the debates after already agreeing to them. Moreover, Biden himself said in an interview after Pelosi's statement that he was going to debate6.

So, it seems unlikely that Biden will back out now. No matter the reasons the Biden campaign might have, or what reasons they would give in public, for withdrawing from the debates now, doing so would support many people's suspicions that Biden is simply not up to it. As a result, he would probably take a hit in support. Of course, if he does debate and performs poorly, it will tend to confirm those suspicions.

Given Biden's age, the vice presidential debate between incumbent Mike Pence and challenger Kamala Harris7, will also be an unusually important one. If Biden's elected, it's more likely than usual that he will either die in office or be incapacitated in some way, in which case Harris will become president.

So, it appears that more is riding on these debates than in a typical election year. The first presidential debate is scheduled for about a month from now8. As usual, I intend not to watch the debates but to read the transcripts after they're over, then comment in a weblog entry within a day or two if I have anything that seems worth saying.


  1. Nicholas Wu, "Pelosi: 'I don't think there should be any debates' between Biden and Trump", USA Today, 8/27/2020.
  2. Benjamin Fearnow, "Supporters Urge Joe Biden Not to Debate Trump, Applaud Hosts for Canceling Over COVID-19 Concerns", Newsweek, 8/1/2020.
  3. "Debate History", The Commission on Presidential Debates, accessed: 8/31/2020.
  4. Alan Schroeder, Presidential Debates: Forty Years of High-Risk TV (2000), pp. 49-50.
  5. "October 21, 1984 Debate Transcript", The Commission on Presidential Debates, 10/21/1984.
  6. Morgan Chalfant, "Biden insists he will debate Trump after Pelosi remarks", The Hill, 8/27/2020.
  7. Kate Sullivan, "Kamala Harris and Mike Pence scheduled to face off at October debate", CNN, 8/12/2020.
  8. "2020 Presidential Debates", The Commission on Presidential Debates, accessed: 8/29/2020.

August 23rd, 2020 (Permalink)

Survey Says

Here's a question for you: What percentage of the population of the United States has died so far with COVID-19?

I'm not expecting you to know the precise answer off the top of your head, nor do I expect you to do research to find out, so just make a guess.

Have you made a guess? Good! A recent survey asked the following question of a thousand adults from each of five countries, including the United States: "How many people in your country have died from coronavirus1?" Those surveyed were asked to answer the question in the form of a percentage. That's where I got the above question.

Before we look at what the survey respondents said and what the correct answer is, let's estimate it rather than just guess. I don't expect to get the exact figure, but an approximation will be good enough. We should be able to come up with an answer within an order of magnitude based on two pieces of information:

  1. Approximately how many Americans with COVID-19 have died to date?

    If you have been paying any attention at all to the American news media recently, you ought to have some idea of this number. So, without peeking at the internet, what do you think it is? You might want to write it down. Remember: we don't need a precise number―we'll look at that later.

  2. About what is the current population of the United States?

    Again, we don't need an exact number, which is impossible to get anyway since it's changing constantly, so don't do any research. Just make your best estimate and we'll look at a more precise answer later. Draw a line and write down your second estimate beneath the first.

To estimate the percentage the question asked for, all you need to do now is to divide the first number by the second, then multiply the result by 100. Since these are big numbers, you might want to use a calculator to do it. When you have finished the estimate, click on the following button to see the answers:

How close was your initial guess? If it was greater than the true percentage by more than an order of magnitude, the survey suggests that you're normal. Unless the survey went very wrong in some way, it appears that the American people greatly over-estimate the severity or prevalence of COVID-19: the mean estimate in the survey was 9%, which is almost 200 times too high4! Moreover, the citizens of all five countries surveyed were off in their estimates in the same direction by two orders of magnitude. So, we can't write this off as simply an American aberration.

How close was your estimate? Hopefully, it was better than your guess and was the right order of magnitude, namely, hundredths of a percentage point. If only the respondents who took the survey had known how to make such an estimate, the resulting answer wouldn't have been so embarrassingly bad.

However, one good result of this survey is that we can see just how poorly informed most citizens of the United States and western Europe are about COVID-19. Perhaps these large errors may be put down to general innumeracy, but given the months of constant news media coverage of coronavirus, why are people not better informed? I think the blame can be apportioned to three institutions:

  1. Government health authorities at the state, national (CDC), and international (WHO) level: I've shown in previous entries that some health agencies have miscommunicated directly to the public, or indirectly through the news media. This miscommunication is not, so far as I've found, the result of outright lies but of failing to tell the full story, thus giving false impressions. Many of these agencies will inundate and confuse you with an avalanche of statistics, but fail to communicate those statistics in a way that puts them into perspective.
  2. Some politicians and public health officials have over-extrapolated numbers of cases and deaths from the available data. Such exaggerated speculation is likely to have contributed to people's exaggerated idea of both the percentage of people who have died from the disease as well as the percentage who have contracted it5.
  3. The news media: Given its tendency to over-report the epidemic to the exclusion of all other news, why is it that people are so misinformed? I think the excessive reporting has itself contributed to the exaggerated fears that so many people have of the virus. In addition, mainstream news outlets have served as a conduit for the misinformation streaming from politicians, government health organizations and officials. The survey results indicate that the news media have failed in their duty to skeptically examine and correct that misinformation, rather than exploiting it for views.

While I wouldn't expect people to know the precise percentage of Americans who have died from COVID-19―I didn't know it myself until I began to research this entry―they should know it's not even close to 9%. Less than 1% would be a good guess, with less than a tenth of a percent an even better one.

It's no wonder that the over-reaction to the coming of the coronavirus continues when people are so badly misinformed about it. When will our politicians, public health authorities, and news media start correcting this misinformation instead of contributing to it?


  1. "COVID-19 Opinion Tracker, Edition 4", Kekst CNC, 7/27/2020, p. 24.
  2. "Cases and Deaths in the U.S.", Centers for Disease Control and Prevention, accessed: 8/23/2020.
  3. "U.S. and World Population Clock", United States Census Bureau, accessed: 8/23/2020.
  4. According to the survey, the American mean estimate was off by 225 times the number of confirmed deaths, but the survey was done over a month ago, when the number of fatalities had just surpassed 130,000. See Note 1, above.
  5. The same survey shows that its respondents also consistently over-estimated the percentage of cases in their country, though not to the same degree as they over-estimated the percentage of deaths. See Note 1, above.

August 12th, 2020 (Updates: 8/13/2020, 8/14/2020, 8/15/2020, 8/16/2020) (Permalink)

Florida's Record-High Death Tolls

We've seen in previous entries that there is often a delay of days to weeks in the reporting of COVID-19 cases from local to state health authorities. The state health department may then release the numbers of cases and deaths to the news media via a press release. When backlogs in reporting from city or county health departments to the state occur, and those cases are finally sent to the state all at once, a large number of cases may be released to the public in a single day1. If the number of cases released is the highest that has been released up till then, the news media may declare it a "record" number2.

For example, check out the following headline and the first couple of sentences beneath it:

Florida reports new daily death toll record

In Florida, a record 191 new fatalities were reported in one day, according to data released by the Florida Department of Health Tuesday morning. The previous one-day record was 173 fatalities, reported on July 23.3

Notice that the headline refers to reporting by the state, and the first sentence of the underlying story says that Florida's Department of Health (FDOH) "reported" the new fatalities "in one day". Similarly, the next sentence says that the "previous one-day record" was "reported". In other words, both of these records are for number of deaths reported in one day, and not necessarily the number who died that day.

As a matter of fact, FDOH's press release for the 28th of last month reported 186 new deaths4, which is not a large discrepancy but what accounts for it? Perhaps the additional five fatalities in the news report came from a subsequent update, or they represent non-resident deaths, since the press release specifically refers to "186 Florida resident deaths related to COVID-19".

So, if the number of deaths reported by the FDOH on a given day is not necessarily the number of fatalities that occurred that day, how can we tell how many did die on that date? In addition to putting out press releases, the FDOH also maintains a COVID-19 Data and Surveillance Dashboard5, which is updated daily. The dashboard includes a chart showing the date of death for residents for about the last month. Beneath that chart occurs the following small-print warning: "Death data often has significant delays in reporting, so data within the past two weeks will be updated frequently." This is the only place on the FDOH's extensive website that I've found such a warning, but it's for this reason that I've waited a fortnight after the above "record" day to consult the dashboard for this entry.

As of today, according to the FDOH chart, there were 135 resident deaths on the 28th, not 191 or even 186. Moreover, on the supposed previous record day of the 23rd there were in fact a non-record 149. When was the actual record number of deaths? As far as the chart goes back, the record-holding day was the 16th with 172 fatalities.

Because of the lag time in reporting to the state health department, it's possible that the above numbers will change in the near future, especially the more recent ones. However, due to these very delays the daily number released by the FDOH does not represent deaths that occurred, but only those that have been reported, on that day. Therefore, "record" numbers in such reports are virtually meaningless, at least if we want to know about the current course of the disease as opposed to the workings of Florida's health surveillance system. The breathless reporting of such "records" is bound to create the false impression that deaths in the state are "surging" or "spiking". FDOH Reported Fatalities

An examination of the reported fatalities from FDOH's press releases shows a distinct wave pattern, with prominent peaks and troughs: see the chart to the right, which shows the last thirty days of reports. The daily fatalities range from a low of 62 on the second of this month to a high of 276 just yesterday. The waves mostly follow a weekly pattern, with one peak a week followed by one trough. Moreover, the peaks tend to occur in midweek, with the troughs occurring on the following Sunday. Two peaks occur on Tuesday―8/4 & 8/11―two on Thursday―7/16 & 7/23―one on Friday―7/31―and one on Saturday―8/8. Three of the troughs occur on Sunday―7/19, 8/2 & 8/9―another is on both Sunday and Monday―7/26-27―and only one is in midweek, specifically, Thursday―8/6.

I can't think of a biological explanation for this wave pattern. Is there some reason why people are less likely to die on Sunday? Rather, it must be an expression of the internal workings of Florida's health bureaucracy: the reporting of deaths is slow over the three-day weekend then is presumably busy catching up in midweek.

As mentioned above, the FDOH has exceeded its prior "record" report of fatalities, with a new record of 276 reported yesterday6. How much of this is the result of Florida's surveillance system suddenly spitting out a backlog of cases? The dashboard is currently showing only 22 deaths yesterday, but that number will surely rise and we won't know for at least two weeks how many actually died that day. As one would expect, the news media are back on the case:

Florida Sees New Daily Coronavirus Death Toll Record

Florida set another record for new daily deaths from the coronavirus on Tuesday when it reported more than 270 deaths. The state's health department reported 276 resident deaths Tuesday….7

Both the headline and first sentence give the false impression that all of these "reported" deaths occurred yesterday, and nowhere in the article is any mention made of a lag time of days or even weeks in such reporting.

Why do news outlets report these misleading and meaningless records? One reason is that the FDOH hasn't done a good job of warning that the numbers cited in the daily press releases do not represent the date of death, nor have they explained that any sudden increases or decreases are probably due to their reporting system rather than any changes in the number of people dying. Another reason is the news media's preference for bad news: good news is no news. Yet another reason is sheer laziness, since it's easier to rewrite a press release than to do actual reporting. Finally, many reporters these days are under-trained and over-worked, which are institutional problems rather than individual ones.

Bad reporting can lead to bad policies. Florida politicians have been under pressure, based on misleading reports such as these, to delay or even backtrack on restarting the state's economy and letting children return to school. Such momentous decisions need to be based on a correct understanding of what the numbers mean rather than on bogus "records".

Update, 8/13/2020: Yesterday, we saw that a supposed record number of deaths from COVID-19 in a single day, namely, the 28th of last month, was really no such thing. The next day, the carnage continued, according to the news media:

Florida reports record-high COVID-19 deaths for the second day in a row…

Florida broke its record for new COVID-19 deaths on Wednesday, the second time this week…. The Florida Department of Health announced 216 new coronavirus deaths on July 29…. The previous daily record was 186 deaths, set the day before.8

As we've seen, the record on the previous day was only a record release of the number of deaths to the public via press release and not one in actual fatalities. The headline above is true since Florida did report 216 new deaths9, but the first sentence of the report beneath it is false, as there were only 132 resident deaths on this day, according to the dashboard10. Far from being a new record number of deaths, this is actually slightly fewer than the previous day. In addition, it is also fewer than the actual record number of deaths set on the 16th of last month.

Not all the reporting of Florida's "surge" was as bad as those examined above, and the news outlets that have done a better job deserve some praise. Some managed to point out that there was a lag time in the reporting of fatalities. However, even then, the reports are usually headed by the typical scary headlines of new "records", with the delays only mentioned as an afterthought. Typical of such reporting is an article which, after several paragraphs of the usual scare-mongering, ends with the following paragraph:

In Florida, when state health officials report a death in the daily update, the number encompasses several days, with actual days of death occurring weeks prior in some cases. The state DOH maintains a chart on its dashboard that shows the actual deadliest days within the last month.11

Well, thanks! In other words, ignore everything you just read. A better job is done in the following article. Here is the headline and the first five sentences in their entirety, but be careful that you don't get whiplash from doing a double-take while reading it:

Florida hit with 216 coronavirus deaths, breaking a record for the second day in a row

Florida's Department of Health on Wednesday confirmed 9,446 additional cases of COVID-19, bringing the stateís known total to 451,423. There were also 216 Florida resident deaths announced, setting another fatality record in the state for the second day in a row. The statewide resident death toll is now at 6,333. The 216 deaths mark the highest single-day Florida resident death toll announced by the Florida Department of Health since the pandemic began, but it does not necessarily mean that every person died in the past 24 hours. In Florida, the deaths announced on a given day could be from several days earlier because the state information does not include the exact date of death.12

Here, at least, we don't have to wait until the last paragraph of the article to find out that the "record" doesn't mean that those 216 people all died on one day, though those who only read the headline―and I often do this myself unless I'm particularly interested in a story―will probably come away with that impression.

Update, 8/14/2020: Florida's record-setting streak continued for another day:

Florida Sets Daily Coronavirus Death Toll Record for Third Consecutive Day

Florida reported more than 250 new deaths from the coronavirus on Thursday, setting another one-day record for new deaths for the third consecutive day. The state's health department reported 253 new deaths from the coronavirus among Florida residents on Thursday…. Thursday's toll broke Wednesday's record of 216 deaths, which had previously broken Tuesday's record of 186.13

This article is the reverse of yesterday's in that the headline is false but its first sentence true. The FDOH did indeed report 253 new fatalities in its press release for the day14, but the dashboard gives almost a hundred less: 15415. This is indeed more than the previous two record days, but still less than the record of 172, which is currently shared by the 16th and 20th of last month.

Update, 8/15/2020: Florida's unlucky streak continued another day:

Fourth day of record-breaking death tolls as Florida counts 257 fatalities16

Once again, the FDOH did "count" 257 deaths this day17, though this was not the day that many of the deaths occurred as the dashboard shows only 15018, which is slightly less than the previous day. This was the last "record-setting" day for this streak, though there was another one eleven days later, as we saw in the original entry.

Update, 8/16/2020: To cap off the four consecutive record-setting days, there's this report:

Florida Coronavirus Deaths Hit Record-High 1,230 In One Week

Florida's COVID-19 fatality toll continues to surge, with weekly deaths setting another dismal record. The state reported 1,230 deaths July 26-Aug. 2―the most in a single week. Florida…shattered its previous weekly fatality record of 872 set July 19-26. Florida authorities said there were a record-breaking 257 deaths statewide Friday, the largest single-day death toll to date. The previous high for daily deaths was 143 on July 17. On Sunday, Florida recorded 62 more deaths…. There was some positive news for Florida, however. The number of deaths reported Sunday was the lowest in nearly three weeks at 62….19

In fact, the FDOH reported 1,245 fatalities for the new record week, and the previous one was 88220. As you might by now expect, the actual number of fatalities for those weeks, according to the dashboard, were 1,157 for 7/19-25 and 1,043 for 7/26-8/1. So, the new "record" actually misrepresented a slight decline in deaths from the previous week.

Amusingly, even the "positive news" at the end of the above excerpt is an artifact of Florida's reporting system. As we saw in the original entry, the FDOH tends to report the lowest number of deaths on Sundays, presumably because fewer people are working then. According to the dashboard, over twice as many people died that day, namely 13221.

Here's an example of the way in which these meaningless records are being used to make political hay:

Trump Says Florida COVID Cases Going Down as State Breaks Death Record

During Thursday's White House coronavirus briefing, Republican President Donald Trump said Florida COVID-19 case numbers are going down even though the state just projected its highest-ever daily toll of deaths. … "The numbers are coming down and coming down very substantially. They're starting to come down in Florida," Trump said during the Wednesday briefing. … However, on Wednesday, Florida's Department of Health reported 216 new COVID-19-related deaths, a new single-day record that followed its previous highest-ever daily death toll of 186 new deaths the day before. As of July 30, the three aforementioned state's COVID-19 daily case numbers have slightly lowered from their highest-ever recent daily tolls. …Florida has averaged 10,250 new cases each day….22

All the above headline is literally saying is that both of those things happened at the same time, but of all the things that happened at that time, why juxtapose those two? Obviously, the point of doing so is to suggest that Trump was wrong in saying that the cases were declining, and the supposed record number of deaths shows it. However, as we've seen, the record was only one in the number of deaths that happened to be made public that day by the FDOH. In fact, the number of deaths had been declining since the middle of last month. Moreover, even if it were a true record of the number of fatalities that day, it wouldn't show that case numbers weren't going down, since cases and deaths are two different statistics, and a peak in deaths tends to follow a peak in cases by a week or two. Finally, just like fatalities, cases in Florida had declined at the end of the month from a peak in its middle, and have continued to decline since23. So, mirabile dictu, Trump was actually right about something.


  1. Spike It, 6/22/2020
  2. Spike it, Again, 7/15/2020
  3. Jon Haworth, Emily Shapiro & Meredith Deliso, "Florida reports new daily death toll record", ABC News, 7/28/2020
  4. "Florida Department of Health Updates New COVID-19 Cases, Announces One Hundred Eighty-Six Deaths Related to COVID-19", Florida Department of Health, 7/28/2020
  5. "Florida's COVID-19 Data and Surveillance Dashboard", Florida Department of Health, Division of Disease Control and Health Protection, accessed: 8/12/2020
  6. "Florida Department of Health Updates New COVID-19 Cases, Announces Two Hundred Seventy-Six Deaths Related to COVID-19", Florida Department of Health, 8/11/2020
  7. Alexa Lardieri, "Florida Sees New Daily Coronavirus Death Toll Record", U.S. News & World Report, 8/11/2020
  8. Charles Davis , "Florida reports record-high COVID-19 deaths for the second day in a row―as the state prepares for a tropical storm", Business Insider, 7/29/2020
  9. "Florida Department of Health Updates New COVID-19 Cases, Announces Two Hundred Sixteen Deaths Related to COVID-19", Florida Department of Health, 7/29/2020
  10. See note 5, above; accessed: 8/13/2020
  11. Tiffini Theisen, "216 newly reported Florida coronavirus deaths is second daily record in a row", Orlando Sentinel, 7/29/2020
  12. Michelle Marchante, "Florida hit with 216 coronavirus deaths, breaking a record for the second day in a row", The Miami Herald, 7/30/2020
  13. Alexa Lardieri, "Florida Sets Daily Coronavirus Death Toll Record for Third Consecutive Day", U. S. News & World Report, 7/30/2020
  14. "Florida Department of Health Updates New COVID-19 Cases, Announces Two Hundred Fifty-Three Deaths Related to COVID-19", Florida Department of Health, 7/30/2020
  15. See note 5, above; accessed: 8/14/2020
  16. Staff Reports, "Fourth day of record-breaking death tolls as Florida counts 257 fatalities", Florida Politics, 7/31/2020
  17. "Florida Department of Health Updates New COVID-19 Cases, Announces Two Hundred Fifty-Seven Deaths Related to COVID-19", Florida Department of Health, 7/31/2020
  18. See note 5, above; accessed: 8/15/2020
  19. Arthur Villasanta, "Florida Coronavirus Deaths Hit Record-High 1,230 In One Week", International Business Times, 8/3/2020
  20. Each of the "weeks" cited is actually eight days, which is not a week―pace the Beatles―so I didn't include the last day listed in either tally.
  21. See note 5, above; accessed: 8/16/2020
  22. Daniel Villarreal, "Trump Says Florida COVID Cases Going Down as State Breaks Death Record", Newsweek, 7/30/2020
  23. See the "New Cases of Residents by Day" chart on the dashboard, just above the "Resident Deaths by Date of Death" chart; accessed: 8/16/2020

Recommended Reading
August 5th, 2020 (Permalink)

Mail Bag

This month's recommended readings all take the form of letters.

Disclaimer: The views expressed in the above readings are solely those of the individuals providing them and do not necessarily reflect the opinions of The Fallacy Files or any of its employees or assignees. Spelling in the Jefferson letter has been modernized. Presented for entertainment and educational purposes only. May cause drowsiness. Do not operate heavy machinery while reading.

August 3rd, 2020 (Permalink)

Prussian Roulette

If you were a spy back in the days when Prussia still existed, and you were captured by the Prussian Secret Service, you would have been subjected to a "game" of Prussian roulette. It wasn't a fun game, at least not for you. The game was like Russian roulette, but with two bullets, since the Prussians prided themselves on being twice as mean as the Russians. It used a six-shot revolver, with two live cartridges inserted into the cylinder in chambers next to each other, and the other four chambers left empty. The cylinder was then closed and spun. The barrel would be held to your head and the trigger pulled.

If you survived the first round of the game, you would be subjected to a second round, but this time you would be given a choice: either the cylinder would be spun once again before pulling the trigger, or it would not be spun.

Assuming that you were a spy captured by the Prussians, which should you have chosen: spin or no spin?

Extra Credit: What would be your chances of surviving a game of Prussian roulette if you chose to spin the cylinder on the second round, and if you chose not to spin it?

August 1st, 2020 (Permalink)

Who Will Fact-Check the Fact-Checkers? (Part 2)

Part 1 examined a factual failure at The New Yorker1, once famed for its rigorous fact-checking. Part 2 is about a supposedly true story by "police abolitionist" Derecka Purnell published by The Atlantic. Here's the operative paragraph:

The first shooting I witnessed was by a cop. I was 12. He was angry that his cousin skipped a sign-in sheet at my neighborhood recreation center. I was teaching my sister how to shoot free throws when the officer stormed in alongside the court, drew his weapon, and shot the boy in the arm. My sister and I hid in the locker room for hours afterward. The officer was back at work the following week.2

However, according to Christopher Bedford of The Federalist, important details of the story were false:

The article's title and call for police abolition remain unchanged, although the story justifying her activism is no longer about (1) a police officer shooting (2) a child (3) without serious consequences, and is about now (1) a private security guard shooting (2) an adult (3) and being charged with assault.3

Purnell subsequently confirmed this by "tweeting":

The shooter was a uniformed private guard with a badge and gun. When we say abolish the police, that includes private police, too.4

The Atlantic went on to correct the story5, thus vitiating whatever rhetorical power the opening anecdote had to support the call for abolishing the police. That the resulting article ends up advocating abolition of the police based on an incident in which a security guard shot someone and then was arrested by the police makes about as much sense as any other argument for abolishing the police. A single anecdote is very weak evidence, even when the anecdote is true, but a false anecdote is evidence of nothing except for the author's bad memory or, perhaps, youthful misunderstanding of what happened.

Why did The Atlantic publish this article? Was it fact-checked before publication? If so, why weren't the falsehoods caught? If not, why not? According to Bedford:

The Atlantic still refuses to share any corroborating evidence or if they did a fact-check on the original story before publication…. When asked if The Atlantic spoke to the victim, spoke to the guard, or acquired a police report, Anna Bross, a vice president of communications at the magazine, replied, "To start, you can find coverage of the incident in local newspapers in 2004." … While still declining to say if the article was fact-checked before it was posted on July 6, and if so by who, Bross emailed that she "will keep an eye out for the significant corrections or updates to your piece(s)," referencing The Federalist investigation that fact-checked the article for them.3

The fact that The Atlantic keeps stone-walling about whether it fact-checked the article indicates that it probably didn't. Like the statistic from the Lepore article examined in part 1, the original anecdote was implausible. We were supposed to believe that a police officer would shoot his young cousin in the arm for not signing in at a recreation center. Then, the cop was back on the job the next week as though nothing had happened. Perhaps this would sound plausible to a thirteen-year-old, but it shouldn't to any adult. So, how did this get past the editors at The Atlantic? Presumably, the story was simply too good to check.

What does Purnell think would happen if the police were abolished? If she would read the Lepore New Yorker article referred to in Part 1 she'd learn a little about what happened before there were police, and why police were established in the first place:

[The] history [of the police] begins in England, in the thirteenth century, when maintaining the king's peace became the duty of an officer of the court called a constable, aided by his watchmen: every male adult could be called on to take a turn walking a ward at night and, if trouble came, to raise a hue and cry. This practice lasted for centuries. (A version endures: George Zimmerman, when he shot and killed Trayvon Martin, in 2012, was serving on his neighborhood watch.) The watch didn't work especially well in England―"The average constable is an ignoramus who knows little or nothing of the law," Blackstone wrote―and it didn't work especially well in England's colonies. Rich men paid poor men to take their turns on the watch, which meant that most watchmen were either very elderly or very poor, and very exhausted from working all day. …

…[U]nlike their British counterparts, American police carried guns, initially their own. … American police carried guns because Americans carried guns, including Americans who lived in parts of the country where they hunted for food and defended their livestock from wild animals, [and] Americans who lived in parts of the country that had no police…. Outside big cities, law-enforcement officers were scarce. … Meanwhile, Americans became vigilantes, especially likely to kill indigenous peoples, and to lynch people of color. … A San Francisco vigilance committee established in 1851 arrested, tried, and hanged people; it boasted a membership in the thousands. An L.A. vigilance committee targeted and lynched Chinese immigrants.6

How does Purnell propose, as in her "tweet", to get rid of private security guards? Would she make them illegal? How will that law be enforced without police? Does she desire a return to vigilantism? We're already seeing this happening in Minneapolis and other cities at the center of the anti-police agitation that she supports7. How would she stop it? Abolishing the police would only return us to vigilance committees and lynch mobs.


  1. Who Will Fact-Check the Fact-Checkers?, 7/30/2020
  2. Derecka Purnell, "How I Became a Police Abolitionist", The Atlantic, 7/6/2020. This is the original, uncorrected version of the story, courtesy of the Internet Archive's Wayback Machine.
  3. Christopher Bedford, "The Atlantic Finally Admits Its Police Abolition Piece Is Based On A False Narrative", The Federalist, 7/21/2020.
  4. Derecka Purnell, "Tweet", Twitter, 7/20/2020.
  5. Derecka Purnell, "How I Became a Police Abolitionist", The Atlantic, 7/20/2020. This is the corrected version.
  6. Jill Lepore, "The Invention of the Police", The New Yorker, 7/13/2020. Warning: Contains the four-letter f-word. No, not "fact".
  7. See, for instance: Leila Fadel, "Armed Neighborhood Groups Form In The Absence Of Police Protection", NPR, 6/3/2020.

Previous Month | RSS/XML | Current | Next Month | Top of Page