Found 15 comments on HN
wyc · 2017-06-23 · Original thread
How to Lie with Statistics is a classic on this:

mjfl · 2017-01-21 · Original thread
Statistics is not trusted because it shouldn't be. It is incredibly easy to lie with statistics. There's even a tutorial:

Statistics is ultimately counting, and therefore is incredibly vulnerable to discretion in choosing "what counts". Take the unemployment rate as one example. When people realize that its not equivalent to "people who do not have a job", how can you complain that they trust statistics less?

Expert authority is in decline because it should be, as there is an increasing body of evidence that experts, from politics to medicine, have almost no advantage in forecasting power than the average person.

Why should "experts" (often just pundits) have any authority when they have consistently demonstrated they deserve very little?

Finally, the political slant of this article, going along with the decried "fake news", blaming the election results on these declines in authority, is pathetic. It's basically an extension of "the other side is filled with stupids" and has no credibility, no matter how you dress it in professional journalistic veneer.

551199 · 2016-11-17 · Original thread
Sure but what if it was never a fact. 'Fat makes you fat' or similar falsification that have real impact to society[0]

We shouldn't attribute everything to malice, but there are clear issues in science:

'Too many of the findings that fill the academic either are the result of shoddy experiments or poor analysis

Last year researchers at one biotech firm, Amgen, found they could reproduce just six of 53 “landmark” studies in cancer research. Earlier, a group at Bayer, a drug company, managed to repeat just a quarter of 67 similarly important papers. A leading computer scientist frets that three-quarters of papers in his subfield are bunk. In 2000-10 roughly 80,000 patients took part in clinical trials based on research that was later retracted because of mistakes or improprieties.

Careerism also encourages exaggeration and the cherry-picking of results. In order to safeguard their exclusivity, the leading journals impose high rejection rates: in excess of 90% of submitted manuscripts. The most striking findings have the greatest chance of making it onto the page.

Conversely, failures to prove a hypothesis are rarely even offered for publication, let alone accepted. “Negative results” now account for only 14% of published papers, down from 30% in 1990. Yet knowing what is false is as important to science as knowing what is true. The failure to report failures means that researchers waste money and effort exploring blind alleys already investigated by other scientists.'[1]

[0] [1]

[2] Alan Sokal - Beyond the Hoax: Science, Philosophy and Culture [3] How to Lie with Statistics - [4] Nassim Taleb 'Incerto',

Unfortunately, sites like snopes and politifact also succumb to the same type of bias as the right-wing sites. See this article where politifact was handed propaganda material from the Clinton foundation and parroted it without doing any actual checking about AIDS drugs the Clinton foundation was funding:

The problem is a lack of education by the consumer. I suggest everyone read this book as an intro on the topic:

Everyone has an agenda, follow the money, and trust no one. Whether it's right-wing like Alex Jones or Left-wing like the Tampa Bay Times a.k.a politifact, you need to be suspicious and do your own research if you want the truth.

pokoleo · 2016-03-27 · Original thread
For more reading on the topic, I'd recommend "How to Lie with Statistics"[0]. It's a short read (144 pages), with mainly tongue-in-cheek instructions on how to mislead.

Most of the usual tricks ('drop the axes', percentage-points, etc) are there, but there are many other, less obvious tricks.

One of the cooler arguments in the book is that it's easy to lean on someone's implicit assumption of volume to modify their understanding.

If you inflate a 15% increase in house spending to look larger than it is, drawing pictures of houses that are 15% wider will make people intuit a 50%[1] increase, despite reading 15%. The author suggests that even if you're incredibly clear with the text surrounding the charts, people still use the charts to understand the scales of change.

[0] (disclaimer: affiliate link) [1] 1.15^3 ~= 1.5

nathancahill · 2014-04-14 · Original thread
I highly recommend reading "How to Lie with Statistics"[0]. Old but still very relevant.


nathancahill · 2014-01-24 · Original thread
How does 365 data points give your conclusions any sort of statistical significance? Snapchat's numbers are allegedly over 30M, Facebook's are over 1.2B and Instagram's are 150M.

I suggest reading "How to Lie with Statistics"[1]


Edit: I guess that sounded a little harsh. No doubt, Snapchat's growth has been explosive and if this was just another blog-spam post about that fact, I'd let it be. But you're selling a statistical software! I'd expect better from people who know what they're doing. Also, "How to Lie with Statistics" is a really good book, should be required reading for anyone dealing with numbers. Did not mean that as an attack on your product.

mathattack · 2013-12-31 · Original thread
I like the story of the graph a lot more than the graph itself.

The two things I would like to see are:

- Per capita debt per person who attended college. (Or perhaps who graduated college) This would answer an implied question of "What if we're just getting more people going to college?"

- The salary legend should start at 0. This would put the relative movement of salary in a more accurate context.

I don't think fixing these changes the story of "The long term cost of college is going up, while the short term benefits are going down." but when I see tricks out of "How to lie with Statistics"[1] my BS detector goes up.


>Yeah, so can lightning, bathtubs, and aspirin. The risk from vaccines is miniscule.

Unlike lightning, risk of harm from vaccines is preventable. One commenter mentioned the Polio vaccine earlier. The OPV does cause a small number of people to develop polio. The matter was paid attention to, OPV use was curtailed in favor of the other vaccine, the problem was solved. The reformulated vaccine retained its effectiveness without further incidences of Polio infection. If it happened today, I would expect people like you tell those first Polio vaccine victims to "Piss off!", "It's good enough!", "You're a fake!", "You have a 'genetic predisposition'" or other hand-waviness, "Adverse effects are rare.", "You probably already had Polio.", etc.

>even smoking marijuana, for that matter,

To paraphrase you [Citation_needed]

>Pretty much everything is more dangerous than vaccines,

Most people take various levels of precaution against those risks, yet many will just let anyone in a lab coat jab them without hesitation. Fine for them, fine for you, but don't expect me to fall in line without more information.

>Your kid is at more risk of SIDS than vaccine complications.

Which is why my kids have been vaccinated (with a subset of the vaccines recommended by the schedule). I just don't jump up out of my chair every time some drug company comes out with a new product to sell. Do you rail against people from countries who follow different vaccine schedules from the one in your country? What do you think of the Japanese, who are relative "vaccine deniers"[2] yet somehow miraculously have the longest life expectancy anywhere on the planet?[3]

> a number of class-action lawsuits were filed against GlaxoSmithKline

>In effect, your claim appears to be without basis.

Vaccine court doesn't protect companies when they knowingly or maliciously harm people, my mistake.

>absurd and damaging to society.

What is absurd and damaging to society is the persistent obfuscation and non-disclosure of clinical trial data, and the failure to conduct proper non-rigged trials. The safety of drugs and medicine is important and it must be credible. There is widespread doubt (among people who pay attention) about the system that is supposed to ensure peoples' safety. Instead of addressing reasonable peoples' reasonable concerns, people such as yourself resort to rudeness, specious character attacks, and other fallacious tactics.

>The actual incidence of problems caused by vaccines is much lower than the incidence of problems incorrectly ascribed to vaccines.

Fact is, you don't know if that is true or not. You don't have any proof of it. You merely want it to be true that anyone who claims injury by a vaccine is a fraud, and a litigious cheat. [Citation needed, indeed.]

>re: note not a gov't org

Give Google a try, you'll get the hang of it.

see also: The Cochrane Collaboration

> Vaccines are by and large rigorously tested before entering the marketplace.

If that were really true for the LYMErix vaccine, the trials would have revealed the large fraction of people[1] who were predisposed to adverse reaction. There are serious problems with drugs qualification testing in the US and EU. Why don't you go read Bad Pharma by Ben Goldacre. here is a preview.

>Is ten thousand people not enough for you? ....

The statistics have to be credible.

And: specifically: " The review showed that reliable evidence on influenza vaccines is thin but there is evidence of widespread manipulation of conclusions and spurious notoriety of the studies. "

>Consider the previous statement about failing to comprehend statistics.

If you believe, based on your statistical prowess that the current system of drug/vaccine certification trials will effectively prevent bad drugs/vaccines from reaching the marketplace, then it is you who has an inadequate grasp of statistics.

>Consider also that your statement is blatantly false:

Or consider that my recollection of the details was mistaken slightly. OPV use was curtailed after VAPP was discovered.

>That means that VAPP is one of the only things that is actually less likely to happen to you than a terrorist attack.

As long as you aren't a starving waif in some developing country (or a starving waif from the US). From the Wiki: In immunodeficient children, the risk of VAPP is almost 7,000 times higher

>I dunno, the existence of people who are willing to spread pseudo-"cautionary tales" based on misrepresented data who can make marketing the vaccine a largely fruitless endeavor seems like it very well might be enough to discourage even a benevolent nonprofit organization. I'm glad we don't have such people around here.

After reading that again, I kind of regret that I wasted any time at all responding to you. If you're the sort of person who thinks a pharmaceutical company makes a favorable comparison to a "benevolent nonprofit organization" then it seems unlikely that you and I will find common ground. You're rude, you're flamebaiting, you've mis-characterized my comments, and your snide innuendo is an offensive character attack.

[1] About one-third of the general population is HLA-DR4+ and risks contracting the arthritic condition when exposed to the vaccine, according to the complaint. The HLA-DR4+ trait is easily detected by a routine blood test; however, SmithKline never recommended that doctors screen for the trait before administering the vaccine, the lawsuit alleges. source:



sytelus · 2012-11-27 · Original thread
Some of these seems to be outright lies (like implicitly labeling X axis for different data points) but other techniques are just well crafted data illusions and covered well in books such as How to Lie with Statistics
capo · 2012-10-03 · Original thread
Weirdly enough the fact that this is a marketing campaign by Microsoft seems to be ignored when considering the numbers in the press release. There is also the curious omission of the percentage of people who found Bing to be “better” than Google.

Any reaction to Bing could have been a result of people being exposed to it for the first time not necessarily it being the “better” option. It also could be that people are having a positive impression simply because it’s not as bad as they imagined it would be, as indicated by this claim: “64% of people were surprised by the quality of Bing’s web search results.”.

Also there is the matter of the ridiculous disclaimer about the features being omitted in the side-by-side as if they aren't integral to the search experience, not to mention that seemingly all queries to Google appear to be originating from Seattle which degrades the quality of local queries by user located elsewhere.

On a related note I recommend reading the How to lie with statistics book, it's required reading especially when outlets copy/paste press releases:

tokenadult · 2011-12-24 · Original thread
The infographic being criticized in the article, an infographic trying to make a point about a public policy position, may not have used the most meaningful fact in the first place. If the issue is cost of imprisonment per inmate per year, then the correct comparison is to the spending per full-time student per year, which at Princeton and several other universities is higher than the billed full list price tuition, because Princeton has other sources of revenue besides tuition.

Spending per full-time student figures are collected by the United States federal government, by law, and are reported on the College Results website maintained by a nonprofit organization.

AFTER EDIT: While doing other things away from my computer, I thought about how the submitted article relates to the culture aspired to here on Hacker News. In February 2009, Paul Graham wrote an article "What I've Learned from Hacker News"

looking back on the first two years of Hacker News. He wrote then, "There are two major types of problems a site like Hacker News needs to avoid: bad stories and bad comments." He thought at that time that the steps Hacker News takes to keep out bad stories have been largely successful, and to this date there haven't been any big technical changes (certainly never downvotes on submissions) to screen out bad stories. The author of the submitted article says, "Think before you link" in her example of an infographic about the problem described in the article, and goes on to say, "So before you pick up that infographic, give it a good, hard look." This is the desired culture here on HN. Early in my 1132 days of participation here on HN, I asked more experienced participants if the expectation here is that links are submitted for comment, even if the submitter disagrees with the link, or if submitting a link is an implicit endorsement that the link has at least minimal quality. The participants who kindly replied to my question overwhelmingly said that I and participants here in general should just submit links that they endorse as worth a read, not crap links to stir up comments of disagreement. I agree with the author of the submitted link here that infographics are too eye-catching and resist efforts at fact-checking, and that is is worthwhile to check the underlying sources and facts before passing on a link to an infographic. Way back in 1954 the author of How to Lie with Statistics

pointed out that some lies about statistics are most easily performed with display graphics. Readers have to be on the lookout for such issues.

sireat · 2010-01-11 · Original thread
Two great books for those who need to re/freshen up their statistics:

I haven't had a chance to check out the manga guide to statistics but that might be a decent introduction, as well.

billswift · 2009-05-27 · Original thread
How to Lie with Statistics ( is a short, enjoyable read. It doesn't tell you how to do statistics, but it gives some warning about common problems.

Get dozens of book recommendations delivered straight to your inbox every Thursday.