Previous Month | RSS/XML | Current | Next Month


July 30th, 2018 (Permalink)

A Once in a Lifetime Puzzle

What occurs once in a second, once in a minute, twice in a day, but only once in a lifetime?


July 23rd, 2018 (Permalink)

New Book: The Smear

Whoever can be trusted with very little can also be trusted with much, and whoever is dishonest with very little will also be dishonest with much.7

I had just begun reading reporter Sharyl Attkisson's latest book, The Smear when, in the first chapter, "Birth of the Modern Smear", I came across the following dubious quote. In discussing Nazi propagandist Joseph Goebbels, Attkisson writes: "…[A]pplicable remarks found within the pages of the Goebbels diaries include: … 'The truth is the greatest enemy of the State.'1"

I'll leave it to you to check Randall Bytwerk's "False Nazi Quotations" page to see the reasons for thinking this alleged quote spurious; it's the first quote on the page2. Anyway, it's not my purpose here to discuss why the quote is probably not Goebbels', but what it's doing in Attkisson's book.

In a previous entry3, I discussed the role of factual statements in fact-checking. Attkisson's claim is certainly a statement, that is, it's either true or false that Goebbels wrote that in his diary. Moreover, it should be possible to search the diaries for this statement, at least theoretically, which means that it is a factual statement. So, this is the type of statement that fact-checkers should check. Unfortunately, there is no citation given for this quote in the book, other than the attribution to the lengthy diaries, which makes it exceedingly difficult to verify4.

The quote is exactly the sort of thing that should have been fact-checked, if not by the author herself then by the publisher. It's not difficult to find out that the correctness of the quote is disputed by experts: all you have to do is cut and paste it into a search engine, and Bytwerk's page on fake Nazi quotations should be one of the first few results. Doing such a search and then reading Bytwerk's page would have taken only a few minutes.5

Attkisson has done some good reporting in the past, but this shows either non-existent or extremely lazy research. Whatever reasons Attkisson herself may have had for making this mistake, where was the publisher? The electronic copy of the book that I have was published by Harper, a reputable and long-established publisher. Apparently, it no longer does even minimal fact-checking on what it publishes, assuming that it ever did.

Now, you may think: big deal! Sure, she made a mistake, but it's trivial. True enough, but it would have also been trivial to check. While in itself the error is small, what effect should it have on our confidence in Attkisson as a reporter and her book? Can we be sure that the big claims she makes are correct? If she, her researchers, or her publisher fail to do a fact-check that takes a minute or two, can they be expected to do so when the research is long and difficult?

This case reinforces a point that I've made before6, but it's worth repeating. In this time of fake news and punditry masquerading as reporting, we need to be willing and able to do our own fact-checking, because many of our institutions are failing to do it.

I was looking forward to reading Attkisson's book, but now I'm not so sure. If she can't be bothered to fact-check her book, I don't think I can be bothered to read it.

Update (8/1/2018): In Attkisson's book, four additional quotes are attributed to Goebbels' diaries. I was a little doubtful about a couple of the others, but I concentrated on the one above because it was all I needed to make my point. Randall Bytwerk has now looked into all of these quotes, and I leave it to you to check out his findings8. I will just add that a pattern seems to be emerging.


  1. Sharyl Attkisson, The Smear: How Shady Political Operatives and Fake News Control What You See, What You Think, and How You Vote (2017), p. 12.
  2. Randall Bytwerk, "False Nazi Quotations", German Propaganda Archive, accessed: 7/23/2018. Last year, I discussed quote 4; see: Who is Adolph Hitler and why does he keep saying these terrible things?, 9/11/2017, and quote 5; see: Passage of Propaganda, 7/28/2017.
  3. Fact Vs. Opinion, 6/22/2018.
  4. There are no footnotes, endnotes, or other citations in the electronic version of Attkisson's book that I have. Notes are not just a scholarly affectation: they make it possible for the reader to check quotes or factual claims for themselves. It's for all practical purposes impossible to prove that Goebbels never wrote this, but the burden of proof is on the quoter to establish that the quote is genuine, and Attkisson has failed to meet this burden.
  5. I checked this today (7/23/2018) in Google, where it was the second result; Bing, where it was first; and DuckDuckGo, where it was also first. Checking all three search engines took less than one minute.
  6. The End of Factchecking, 5/17/2014.
  7. Luke 16:10 (NIV), added: 7/28/2018. My thanks to a pseudonymous reader who reminded me of this saying.
  8. Randall Bytwerk, "Sharyl Attkisson Uses Dubious Quotations", Goebbels Didn't Say It, 8/1/2018.

July 6th, 2018 (Permalink)

Poll Watch: Is President Trump more popular than the FBI?

According to the following headline, he is:

Trump Approval Jumps As Attacks From Critics Backfire;
Trump Now More Popular Than FBI: IBD/TIPP Poll1

The July poll found that 41% approve of the job Trump is doing. That's the first time he's been above 40% since March 2017, and just 1 point below his highest approval rating of 42% during his first month in office.1

So, an alternative headline for this story would be:

Trump's Approval Rating Unchanged Since Taking Office

But you won't be seeing that anywhere but here.

Here's what supports the headline's comparison of Trump's approval rating with that of the FBI: "The poll found that just 39% have a favorable view of the FBI1." So, the difference between Trump's approval and that of the FBI is two percentage points. Most modern public opinion polls have a margin of error (MoE) of at least 3 percentage points, so this is not a significant difference. A more accurate appraisal of this result would be that Trump and the FBI are about equally unpopular. At the end of the article, where they always hide this information, you can read5:

Methodology: IBD/TIPP conducted the July poll from June 21 to June 29. It includes responses from 900 people nationwide, who were asked questions by live interviewers on phones. The poll's margin of error is +/-3.3 percentage points.1

This is actually a slightly small sample for a national public opinion poll, and the difference between the President's and the FBI's approval ratings is well within the margin of error2. There's another logical problem with the reporting of this poll, the first sentence of which reads:

Increasingly bitter and aggressive attacks on President Trump managed to bolster his approval rating, which climbed 5 points to reach the second highest level of his presidency, the latest IBD/TIPP Poll finds.1

This sentence claims that attacks on Trump served to "bolster"3 the approval rating, meaning that they supported it, which is a causal relation. However, the most that we know from this poll is that Trump's approval rating increased 5 points from the previous IBD/TIPP poll, and the attacks in question presumably happened either before or during the time that the poll was taken. Of course, it's possible that the attacks may have actually backfired, but it's also possible that Trump's approval rating would have increased even more if it hadn't been for those criticisms. Also, it's possible that the 5 point change was simply a random fluctuation. So, inferring causation in this case commits a causal fallacy4.

This is one of those polling reports in which, if you remove all the exaggeration, speculation, innumeracy, and illogicality, nothing much is left.


  1. John Merline, "Trump Approval Jumps As Attacks From Critics Backfire; Trump Now More Popular Than FBI: IBD/TIPP Poll", Investor's Business Daily, 7/2/2018.
  2. See: How to Read a Poll: Margin of Error Errors.
  3. A bolster is either a long cylindrical pillow that is used to support the human body, or a structural part of a building that provides support. So, as a verb, to bolster means to support something.
  4. See: Causal Fallacy.
  5. I originally wrote here that the difference between Trump's approval rating and that of the FBI was only one percentage point, but it was two. The paragraph has been corrected.

Solution to a Once in a Lifetime Puzzle: The letter "a".

Previous Month | RSS/XML | Current | Next Month