How to Lie with Statistics
Both charming and informative about how statistics are misused. Published long ago, but the tricks haven't changed.
Stats for nerds
63recommend
180mentions
126people
2007–2026years
53% recommend43% neutral4% critical
Mentions per year
What Hacker News says
- There are a number of fairly good books on statistics. One of the best is How To Lie With Statistics by Darrell Huff which was written in 1954. There is also the Cartoon Guide to Statistics by Larry Gonick and Woolcott Smith. The problem most statistics books is that they don't focus on what you can do with statistics. It is important if you are a gambler or want to understand artificial intelligence. Most statistics books are super dry and full of bad examples. I would focus on books like The Theory of Gambling and Statistical Logic. It becomes much more interesting when it is applied.
- Well, if we're telling stats jokes...
A statistician is a man who with his head in the freezer and his feet in the oven can say "On the whole I feel perfectly normal."
Recommended reading: 'How to lie with statistics' - Huff, 1954
:)
- What did they write that article with?
The year is 2026. The unemployment rate just printed 4.28%, AI capex is 2% of GDP (650bn), AI adjacent commodities are up 65% since Jan-23 and approximately 2,800 data centers are planned for construction in the US. In spite of the current displacement narrative – job postings for software engineers are rising rapidly, up 11% YoY. ... We wrote last week that we see the near-term dynamics around the AI capex story as inflationary, but given markets are focused on the forward narrative, we outline a more constructive take on the end state below. Before that, however, it’s worth reflecting that the imminent disintermediation narrative rests on the speed of diffusion.
The chart "Job Postings For Software Engineers Are Rapidly Rising" seems to show a rise from 65 to 71 for "Indeed job postings" from October 2025 to March 2025. That's a 9% increase. Then they inflate that by extrapolating it to a year. The graph exaggerates the change by depressing the zero line to way off the bottom and expanding the scale. This could just be noise.
The chart "Adoption Rate of Generative AI at Work and Home versus the Rate for Other Technologies" has one (1) data point for Generative AI.
This article bashes some iffy numbers into supporting their narrative.
Suggested reading: [1]
[1] https://en.wikipedia.org/wiki/How_to_Lie_with_Statistics
- there is a good book: 'how to lie with statistics'
- True. Math is drier than physics. But Terrence Tao is well known, and there are a lot of nerds on the planet. The last popular math book I read wasn't good (Humble Pi; a Comedy of Math Errors), but I did enjoy e.g. How to Lie with Statistics, and, to a lesser extent, Innumeracy. I would buy Tao's book, probably.
- I believe you write in good faith, and that you sincerely and non-agressively hold your opinion, and that you believe you don't lack a well known piece of information.
But first let me quote a short piece of text, and later in the comment I will reveal where it comes from.
"After World War II, colonial power was increasingly frowned upon on the international stage. To ease pressure from the United Nations, Denmark decided to reclassify Greenland, not as a colony, but as a region. A new Status that required Denmark to guarantee EQUAL LIVING STANDARDS for both Greenlanders and Danes."
Hold on to this "EQUAL LIVING STANDARDS".
So now on towards the poll.
Back when I was studying physics, one of the courses was statistics. Now statistics in physics or mathematical courses is very different from "statistics" in applied / social / political sciences where students are merely required to execute a procedure, like linear regression to fit a line, or the steps to calculate average and variance, ... Those are fixed formulas without rearranging terms and applying mathematical deduction to statistical statements. One can't fully grok statistics in this light form, it needs more rigorous foundations, only then can students learn to derive their own original conclusions in a correct manner and be able to see through the honest mistakes or manipulations of statistical results by others. The professor recommended a booklet called "How to Lie with Statistics". Of course the goal of the book is NOT to breed dishonest statisticians, but to show the myriad of ways statistical results are depicted and phrased to convey intentionally convey an incorrect impression or conclusion, so that we can detect and see through it.
One of the classic things is for example the distribution of top classes: consider mortality rates for different afflictions, lets pretend we buy into the mono-causal paradigm, so tree like, not DAG like. Then if some entity is embarrassed about the top entry, you can just split it up in similarily balanced subcases (instead of a category cancer, splitting it up into all the different kinds of cancer might result in say cardiovascular diseases becoming the top category, simply by splitting up the top class. (My example is arbitrary, I care naught about top mortality, personally).
A false dilemma (false trilemma etc.) is when all the options combined don't form the universe of possibilities, like "would you prefer pestilence or cholera"?
Please take a careful look at the actual poll options [0]:
1. I want independence unconditionally, regardless of the impact on the standard of living
2. I want independence, even if it would have a major negative impact on the standard of living
3. I want independence, even if it would have a small negative impact on my standard of living
4. I only want independence if it doesn't have a negative impact on my standard of living
5. I don't want independence
6. Don't know
It's almost like some Dane made up the vote-able categories and decided to troll the Greenlanders with a reference to the broken promise: LIVING STANDARDS ?!? Some Good Old forced contraception foisted of as the required EQUAL LIVING STANDARDS between Danes and Greenlanders ?!!
So we can classify already: Don't Know (option 6: 9%) vs Know (presumably options 1 through 5: 91% claim to know what they want), so far so good since we have mutually exclusive but exhaustive split.
Now consider the universe of possibilities for those who Know:
Those who know they want independence (from Denmark; options 1 through 4: 84% of all respondents) and those who know they don't want independence (from Denmark; option 5: 9% of all respondents)
So far so good.
Those who want independence (from Denmark) unconditionally (option 1: 18% of all respondents) and those who want independence (from Denmark) conditionally (option 2 through 4: 66%)
Here it gets vague because the boundaries one is asked to get classified in (divide and conquer style) are subjective: on condition there is no "major", "small" or "negative" impact on standard of living.
Is "negative impact" more or less negative than "small negative impact"? I want to see HN commenters discuss if "negative impact" is better or worse than "small negative impact".
This is just non-quantitative gerrymandering.
But let's ignore the gerrymandering: the phrasing is not neutral, as if it is a given there will be negative impact on standards of living!
Imagine the poll stated not the above but:
1. "I want independence unconditionally, regardless if the Danes perform a new round of population control as a goodbye present for old times sake"
2. "I want independence conditionally, regardless if the Danes perform a new major round of population control as a goodbye present for old times sake"
3. "I want independence unconditionally, regardless if the Danes perform a new small round of population control as a goodbye present for old times sake"
4. "I want independence unconditionally, regardless if the Danes perform a new round of population control as a goodbye present for old times sake"
It would be the exact same logical fallacy, but probably with different results, thousands of women (and men) would keelhaul their nearest Danish officials under the nearest ice shelf.
It's just insulting for an (unverifiable) poll to pull these tricks, especially if the poll was co-organized by a Danish newspaper.
> The poll, which was carried out by Verian on behalf of Danish newspaper Berlingske ...
Something else that is insulting: I saw pictures of immense crowds protesting Trump's comments, and read the number of protesters involved: practically the population count of whole Greenland... until I saw the fine print: the numbers were for a protest in Denmark, not Greenland!
Let people speak for themselves, and don't gerrymander polls, its just doubly insulting, and shows that the colonial mentality is still present, sigh!
that power goes both ways, what happens if the Greenland population demands the full list of doctors involved, what type of doctors: military or civilian?, their extradition for legal proceedings on Greenland soil, the confiscation of their pension funds, ... the whole shebang, or else --- who knows they might become a state joining a Union of States, perhaps EU perhaps US. The US has a similar history, from a similar time frame, but the Danish government took a remarkably longer time to even acknowledge what happened.
source:
[0] https://www.euractiv.com/news/virtually-no-greenlander-wants...
[1] https://www.dw.com/en/denmark-apologizes-for-abuse-of-greenl...
Check this documentary (about 30 minutes), horrendous crimes. And then "apologizing", apologizing is when all forms of help have been exhausted, instead of apologizing reveal the lists of doctors, so the Greenlanders can question them, who they got commands from, and were those people got their instructions from, extraditions, confiscation of their pension funds (think about it: having been raped by the doctor, or sedated (another crime if for non medical reasons). The normal order is acknowledge, then help, help, help, and only when all forms of help have been exhausted, apologize.
And Europe is angry how Trump plays the realpolitik game, but by not insisting a Greenland run referendum, but instead backing Denmark, they are playing the realpolitik game just as well, you know "maintaining good relations"
Recommended viewing (30 minutes), its where the quote comes from:
