Status
Available
Call number
Library's review
Indeholder "Indledning", "I. Man tager en stikprøve", "II. Man tager det kønneste tal", "III. Man tager det ikke så nøje - man tager det alt for nøje", "IV. Man tager en myg og kalder den elefant", "V. Man tager en saks", "VI. Man tager på", "VII. Man tager et uvedkommende tal", "VIII. Man
"Indledning" handler om at dette måske nok er et kursus i at kigge på låse og nøgler, men ikke beregnet til tyveknægte, der nok godt kan tricksene i forvejen.
"I. Man tager en stikprøve" handler om at sige at Yale-kandidater tjener 25111 dollars om året, når tallet i virkeligheden bygger på en lille stikprøve.
"II. Man tager det kønneste tal" handler om middeltal, median eller typetal, fx for indkomster i et boligområde.
"III. Man tager det ikke så nøje - man tager det alt for nøje" handler om skamridning af ellers korrekte resultater, fx ved ikke at medtage usikkerhedsintervaller. Lidt om Kinsey og Gesell.
"IV. Man tager en myg og kalder den elefant" handler om overdrevne forskelle.
"V. Man tager en saks" handler om først at beskære akserne og så skalere dem og så evt helt fjerne skalaen.
"VI. Man tager på" handler om at bruge højden af en pengesæk som skala og derved snyde øjet til at tro at forskellen på to tal fx er en faktor 8 i stedet for en faktor 2.
"VII. Man tager et uvedkommende tal" handler om at besvare andre spørgsmål, når nu det stillede ikke kan besvares som ønsket. Oplys fx at halsmedicinen kan dræbe alle bakterier i et reagensglas og lad være med at fortælle at det er i en koncentration ingen vil finde på at bruge indvortes.
"VIII. Man tager Dem ved næsen" handler om korrelationer, indbildte såvel som virkelige. Fx regn indtil en vis mængde er godt for høsten. Mere end det, så bliver høsten mindre. Korrelation beviser ikke nogen årsagssammenhæng.
"IX. Man tager et kursus i svindel" handler om at noget statistik må tilskrives ond mening snarere end uvidenhed. Fx et kort, der viser hvilke delstater i USA der skal til for at dække forbundsregeringens forbrug. Taget fra vest, bliver det ca halvdelen af USA. Taget fra øst, er et par af småstaterne nok.
"X. Vi tager tyren ved hornene" handler om at stille spørgsmål til enhver statistik. 1: Hvem siger det? 2. Hvor ved de det fra? 3. Mangler der noget? 4. Holder man sig til sagen? 5. Lyder det rimeligt? Bogen slutter med et citat fra Mark Twain om at Missisippifloden er blevet 364 kilometer kortere i løbet af 176 år, så om 742 år vil den være helt væk!
Meget morsom bog, som jeg læste dagen efter et kommunalvalg i Danmark, hvor DR bragte exitpolls fra ePinion, der var helt hen i vejret - dvs lige til bogens kapitel et. Og på side 99 står det ultrakort: en procentberegning på grundlag af et lille antal er næsten altid vildledende.
Bogens eksempler er fra 1940erne og 1950erne og som sådan forældede, men man kan bare åbne dagens avis og se friske eksempler.
Show More
tager dem ved næsen", "IX. Man tager et kursus i svindel", "X. Vi tager tyren ved hornene"."Indledning" handler om at dette måske nok er et kursus i at kigge på låse og nøgler, men ikke beregnet til tyveknægte, der nok godt kan tricksene i forvejen.
"I. Man tager en stikprøve" handler om at sige at Yale-kandidater tjener 25111 dollars om året, når tallet i virkeligheden bygger på en lille stikprøve.
"II. Man tager det kønneste tal" handler om middeltal, median eller typetal, fx for indkomster i et boligområde.
"III. Man tager det ikke så nøje - man tager det alt for nøje" handler om skamridning af ellers korrekte resultater, fx ved ikke at medtage usikkerhedsintervaller. Lidt om Kinsey og Gesell.
"IV. Man tager en myg og kalder den elefant" handler om overdrevne forskelle.
"V. Man tager en saks" handler om først at beskære akserne og så skalere dem og så evt helt fjerne skalaen.
"VI. Man tager på" handler om at bruge højden af en pengesæk som skala og derved snyde øjet til at tro at forskellen på to tal fx er en faktor 8 i stedet for en faktor 2.
"VII. Man tager et uvedkommende tal" handler om at besvare andre spørgsmål, når nu det stillede ikke kan besvares som ønsket. Oplys fx at halsmedicinen kan dræbe alle bakterier i et reagensglas og lad være med at fortælle at det er i en koncentration ingen vil finde på at bruge indvortes.
"VIII. Man tager Dem ved næsen" handler om korrelationer, indbildte såvel som virkelige. Fx regn indtil en vis mængde er godt for høsten. Mere end det, så bliver høsten mindre. Korrelation beviser ikke nogen årsagssammenhæng.
"IX. Man tager et kursus i svindel" handler om at noget statistik må tilskrives ond mening snarere end uvidenhed. Fx et kort, der viser hvilke delstater i USA der skal til for at dække forbundsregeringens forbrug. Taget fra vest, bliver det ca halvdelen af USA. Taget fra øst, er et par af småstaterne nok.
"X. Vi tager tyren ved hornene" handler om at stille spørgsmål til enhver statistik. 1: Hvem siger det? 2. Hvor ved de det fra? 3. Mangler der noget? 4. Holder man sig til sagen? 5. Lyder det rimeligt? Bogen slutter med et citat fra Mark Twain om at Missisippifloden er blevet 364 kilometer kortere i løbet af 176 år, så om 742 år vil den være helt væk!
Meget morsom bog, som jeg læste dagen efter et kommunalvalg i Danmark, hvor DR bragte exitpolls fra ePinion, der var helt hen i vejret - dvs lige til bogens kapitel et. Og på side 99 står det ultrakort: en procentberegning på grundlag af et lille antal er næsten altid vildledende.
Bogens eksempler er fra 1940erne og 1950erne og som sådan forældede, men man kan bare åbne dagens avis og se friske eksempler.
Show Less
Publication
Vinten, 1964.
Description
Darrell Huff runs the gamut of every popularly used type of statistic, probes such things as the sample study, the tabulation method, the interview technique, or the way the results are derived from the figures, and points up the countless number of dodges which are used to fool rather than inform.
User reviews
LibraryThing member reading_fox
A useful guide to looking carefully at how data is presented to you in popular media and how the conclusions are then drawn. More journalists should read this to avoid perpetuating the same errors. Editors should read it even more carefully. It is of course a guide to avoiding lies through the use
Show More
of statistics. Very non-mathy and easily understandable, even if you didn't get any maths qualifiactions. Show Less
LibraryThing member slothman
An excellent, easily accessible book that explains common techniques for using statistics to deceive the unwary. The examples are dated-- there's a bit of culture shock in seeing discussions of income from half a century ago!-- but quite clear. The book has lost none of its relevance since it was
Show More
written. Show Less
LibraryThing member AlexTheHunn
This old book is one that endures despite its age. It is a humorous look at how statistics can be used, manipulated, and twisted to say just about anything that you want. In the larger sense, it is important to understand how facts can be spun to represent or create the reality you want to portray.
LibraryThing member BraveKelso
A classic. A great guide to understanding statistical claims in journalism, advertising and politics.
LibraryThing member jocraddock
Easily the most entertaining introduction to statistics and critical thinking I've come across. (What's my experience rate, you ask? You must have read the book!)
While many examples are dated, they remain relevant and the lessons on how to separate the wheat from the chaff are vital. This should be
While many examples are dated, they remain relevant and the lessons on how to separate the wheat from the chaff are vital. This should be
Show More
required reading, but that may make it even more inaccessible. Let's do a study on that! Show Less
LibraryThing member fdholt
How to lie with statistics has become a classic since its publication almost sixty years ago. In an easy to read style, author Darrell Huff explains basic statistics and how results can be skewed by using different statistical techniques and graphs. He tells the reader how to get the best
The chapter on averages was exceptionally clear with the explanations of mean, median and mode. I now know what stasticulating means and how to do it or recognize it. In fact, I will not look at statistics in any newspaper or journal article again without questioning the methods and results.
My copy is the original edition with cartoon drawings by Irving Geis. The examples are from the fifties and seem terribly outdated. However, the important information is not outdated and is still relevant. If you need to understand how statistics work and how they can be manipulated for any purpose, this book is just right.
Show More
advantages while, at the same time, he warns the consumer what may be happening and why things look so “bad” or “good.”The chapter on averages was exceptionally clear with the explanations of mean, median and mode. I now know what stasticulating means and how to do it or recognize it. In fact, I will not look at statistics in any newspaper or journal article again without questioning the methods and results.
My copy is the original edition with cartoon drawings by Irving Geis. The examples are from the fifties and seem terribly outdated. However, the important information is not outdated and is still relevant. If you need to understand how statistics work and how they can be manipulated for any purpose, this book is just right.
Show Less
LibraryThing member everfresh1
Somewhat outdated, somewhat still relevant. Scientific statistics became more reliable, usually providing margin of error and similar parameters. In terms of advertisement/media not much has changed. It was amusing to read about some artifacts from 60 years ago.
LibraryThing member masyukun
This book illustrates common ways to distort statistics. The most valuable part of this book is the last chapter, in which the author arms the reader with a series of questions to help him think critically in the face of authoritative-sounding numbers. This book is funny, educational, and should be
Show More
required reading for every critical thinker. Show Less
LibraryThing member lerned
Great little book, and entertaining too, statistically improbable anyone could make statistics amusing.
LibraryThing member Devil_llama
A reissue of the 1954 classic, the author details many of the mistakes, either deliberate or accidental, that people make when deploying statistics in support of their pet cause. A handy corrective to the current culture of statistical malapropism, and a valuable tool to arm yourself against the
Show More
barrage of bogus statistics and misinterpreted data that assail us every day. Show Less
LibraryThing member statethatiamin
Don't believe anything!
This fast little read does an amazing job uncovering the methods of presenting misleading statistics. This book was first published in 1954, but it reads as fresh as ever. People with axes to grind have been employing the same subversive tactics since statistics have been
This fast little read does an amazing job uncovering the methods of presenting misleading statistics. This book was first published in 1954, but it reads as fresh as ever. People with axes to grind have been employing the same subversive tactics since statistics have been
Show More
popularized. Furthermore, this book is often funny, hilarious even, particularly with the illustrations. The "claim to fame" statistic on the cover is an almost certainly deliberate illustration of ironic contempt for misleading statistics. After all, just because "Twighlight" was a bigger seller than "Anna Karenina" doesn't make it better. Show Less
LibraryThing member TLCrawford
For a book to remain in print for fifty years it must be good. This one was originally published in 1954 and, as far as I can tell, has been in print ever since. A book less than 150 pages long, generously seeded with amusing cartoons is not what you would expect to find on a graduate school
Huff is careful to spread the blame for lying statistics widely, overeager researchers, poor information gathering by statisticians, advertising people willing to apply lipstick of any color to their pig, journalists looking for a marketable story. The fact that most of these lies are “true” is not ignored. For me the most memorable story he uses to make this clear is the restauranteur who explains his rabbit-burger is 50% rabbit, he mixes it in a 1 to 1 ratio with horse-meat. One rabbit to one horse.
After nine chapters of explaining how easy it is for statistics, charts, graphs, and percentages to lie the last chapter makes a serious attempt to explaining how we can avoid being lied to by asking a few simple questions like, who says so, how does he know, what’s missing, and does it make sense. As Huff points out it is important to be able to detect these lies, not just because of misleading advertisements but because we have elections every few years.
As an amateur historian who is just a few years younger than this book I have to admit I enjoyed the window into the past that the many cartoons offered. Yes, we really dressed and smoked like that. The books age was a little disconcerting when Huff dissected an article about the income of the “average” Yale graduate. Going to Yale hardly seemed worth the $25,000 income it offered until I ran it through an inflation calculator, then it made sense. This book is one of the most informative and fun books I have read in a long, long time. It was informative not because I know nothing about statistics, I do, it was informative because nether of the classes I have taken on statistics covered how easy it is to miss-use or misunderstand exactly what it is the numbers say. If you do not like being lied to, consider reading this book.
Show More
reading list but that is exactly where I learned about this one. Darrell Huff and illustrator Irving Geis produced a little marvel with their book “How to Lie with Statistics”. As Huff points out early in the book a cat-burglar who writes a how-to memoir in prison does not do it for other cat-buglers. They already know how to burgle. The intended audience is people who do not want to be burgled, or, in the case of this book, lied to.Huff is careful to spread the blame for lying statistics widely, overeager researchers, poor information gathering by statisticians, advertising people willing to apply lipstick of any color to their pig, journalists looking for a marketable story. The fact that most of these lies are “true” is not ignored. For me the most memorable story he uses to make this clear is the restauranteur who explains his rabbit-burger is 50% rabbit, he mixes it in a 1 to 1 ratio with horse-meat. One rabbit to one horse.
After nine chapters of explaining how easy it is for statistics, charts, graphs, and percentages to lie the last chapter makes a serious attempt to explaining how we can avoid being lied to by asking a few simple questions like, who says so, how does he know, what’s missing, and does it make sense. As Huff points out it is important to be able to detect these lies, not just because of misleading advertisements but because we have elections every few years.
As an amateur historian who is just a few years younger than this book I have to admit I enjoyed the window into the past that the many cartoons offered. Yes, we really dressed and smoked like that. The books age was a little disconcerting when Huff dissected an article about the income of the “average” Yale graduate. Going to Yale hardly seemed worth the $25,000 income it offered until I ran it through an inflation calculator, then it made sense. This book is one of the most informative and fun books I have read in a long, long time. It was informative not because I know nothing about statistics, I do, it was informative because nether of the classes I have taken on statistics covered how easy it is to miss-use or misunderstand exactly what it is the numbers say. If you do not like being lied to, consider reading this book.
Show Less
LibraryThing member kgeorge
This book explains how numbers and statistics can be manipulated to reflect data results in one's favor. The writing is easy to understand and the author uses relevant examples. However, since this book was originally written in 50s, the examples are a little out dated. Nevertheless, a teacher
Show More
could find comparative examples using recent situations and events. This would be a great book to use when talking to students about the importance of displaying their data accurately. It could also be used in a history class when talking about politics and election polls. Show Less
LibraryThing member Gail.C.Bull
I received this book as a required text for a first year university writing class, but it is invaluable to anyone who wants to understand how writers, politicians, and the media manipulate statistics to support their own point of view.
LibraryThing member MartinBodek
Now I know, and so should you, but you probably won't read the book, so I've done it for you.
LibraryThing member KingRat
When I saw How to Lie with Statistics in the free books boxes I jut had to grab it. This is the grand-daddy of line of books about bias in statistics which lead to such descendants as John Allen Paulos’ A Mathematician Reads the Newspaper. I’s a short read that uses humor to impress the subject
Show More
on the population rather than stridency. Show Less
LibraryThing member danrk
What catches you first is its title. What keeps you in it are the examples, humor, and illustrations. Yes, it is somewhat dated, but certainly not outdated and is still relevant -- possibly more than ever with the lifetimes-worth of content on the web. Reading this book will help you sniff out the
The main theme is fairly simple: don't take a number at face value -- there's chicanery afoot. Yes, people want to win arguments and make gains. It turns out numbers help with this. Even if there isn't an element of chicanery (the word's used a lot in the book), clumsiness or carelessness can still lead to false conclusions, and each chapter describes a mischievous tactic for twisting a number to come to those conclusions.
There aren't too many pages and the lessons are invaluable. Take a few hours to read this and keep the lessons in your frontal lobe when you meet numbers supporting conclusions.
Show More
bad stuff so you can focus on the information that counts.The main theme is fairly simple: don't take a number at face value -- there's chicanery afoot. Yes, people want to win arguments and make gains. It turns out numbers help with this. Even if there isn't an element of chicanery (the word's used a lot in the book), clumsiness or carelessness can still lead to false conclusions, and each chapter describes a mischievous tactic for twisting a number to come to those conclusions.
There aren't too many pages and the lessons are invaluable. Take a few hours to read this and keep the lessons in your frontal lobe when you meet numbers supporting conclusions.
Show Less
LibraryThing member trilliams
A good cursory introduction to how easy it is to send wrong messages through statistics. A bit unnerving how a lot of this is still applicable. Don't expect to learn Inception-level skills.
LibraryThing member KamGeb
I picked up this book because a statistics professor I loved read about it. I found my statistics professor explained the ideas better than the book. I also found the book VERY dated. I was expecting an amazing book, and was disappointed.
LibraryThing member kaitanya64
Absolutely necessary classic for anyone who values logic and commonsense.
LibraryThing member trile1000
Short and interesting primer on how to understand the weaknesses behind statistics. We are thrown statistics in one shape or another all the time but what do they actually mean? What is the average? And which "average"? This book shows how statistics can be manipulated and teaches you to be
Show More
skeptical. You learn how to question the conclusions, the graphs, and the data. Concise, practical, and entertaining read. Though the book was published a while ago, the lessons are still relevant today. Would recommend for those who want to be appreciate the subtleties of statistics or be more aware of how numbers can be used to deceive. Show Less
LibraryThing member Daniel.Malcor
Very good read. Should be required before anyone is allowed to read or write the "News". Next question is will they use it to write better or lie better?
LibraryThing member andycyca
Sixty years after its first edition, this book is still true to its purpose: how to catch a lot of tricks, half-truths and purposeful omissions in everyday statistics.
Written for the layman and useful for the specialist, knowledge of these "tips" is more useful than ever in our information
Written for the layman and useful for the specialist, knowledge of these "tips" is more useful than ever in our information
Show More
overloaded society. Do yourself a favor and get a copy. Show Less
LibraryThing member EricCostello
Classic of its kind that hasn't aged a bit since it came out many years ago. A very practical guide to how statistics can be misused, either intentionally or not. A must-read for anyone exposed to statistics on a regular basis -- which means anyone that reads a newspaper or magazine, or watches
Show More
television! Show Less
LibraryThing member elam11
This was entertaining, and I think it could serve as a good introductory text to thinking skeptically about information presented as fact (information literacy skills ahoy!). If you're already familiar with the main deceptive practices for information reporting, you're not going to learn anything
However, it was informative by implicitly revealing some of the concerns of yesteryear. You know, irrelevant things like electoral college votes, toothpaste promises, the assessment of education, and so. :)
Show More
new and different. However, it was informative by implicitly revealing some of the concerns of yesteryear. You know, irrelevant things like electoral college votes, toothpaste promises, the assessment of education, and so. :)
Show Less
Subjects
Awards
Language
Original language
English
Original publication date
1954
Physical description
126 p.; 17.9 cm
Local notes
Omslag: Irving Geis
Omslaget viser en graf med to kurver på og en mand, der fejer nogle tal ind under gulvtæppet
Indskannet omslag - N650U - 150 dpi
Oversat fra amerikansk "How to Lie with Statistics" af Jens Munk-Madsen
Illustreret af Irving Geis
Side 51: Man blæser på den gode gamle regel, at en forskel kun er en forskel, hvis den gør en forskel.
Side 70: Hvis De vil være lige så sludrevorn som artiklens forfatter, kan De bevise at trafikrisikoen er større i klart vejr end i tåge, for der sker færre ulykker i tåge. Det er nemlig kun tåget vejr en mindre del af årets dage. Ikke desto mindre kan tåge være temmelig risikabel for trafikanterne.
Side 79: To ure går nøjagtig ens. Når uret a viser hel, slår b. Får uret a uret b til at slå?
Side 92: Så længe det er regelen at de fejl der forekommer, går til samme side hver gang, er det temmelig søgt at forklare dem som bommerter eller uheld.
Side 96: Dette ejendommelige stykke statistik overdriver på to måder. For det første går man ud fra det aritmetiske gennemsnit i stedet for midttallet, det har vi tidligere været inde på. For det andet bilder man sig tilsyneladende ind at en families indtægt bestemmes af familiemedlemmernes antal. Jeg har selv fire børn, og ville være særdeles tilfreds hvis det skulle blive indrettet sådan, men det tør jeg ikke håbe på. Familier med fire medlemmer er ingenlunde generelt bedre stillet end familier på to.
Side 96: .. det tjener til påmindelse om hvor lidt tiltro et gennemsnitstal er værd, når beregningsmåden ikke er oplyst.
Side 98: En procentberegning på grundlag af et lille antal er næsten altid vildledende. Antallet i sig selv er mere oplysende. Og når procentangivelsen har decimaler med, er det ikke bare pjattet, det nærmer sig fup.
Side 103: Hvis man ganske frimodigt lægger procentuelle stigninger sammen, ender man i det rene vanvid. Tænk Dem at De køber tyve forskellige varer i dag, og efter et års forløb viser det sig at hver af varerne er steget fem procent i pris. Fem procent lagt sammen tyve gange er 100 procent; prisniveauet er altså steget til det dobbelte på et år. Sikke noget sludder.
Side 106: Det kan undertiden være en hel bedrift at påvise at et kontroversielt emne kan anskues anderledes end i sort og hvidt.
Side 109: Konklusionen må blive at der er al mulig grund til at granske statistisk materiale i bøger, blade, og annoncer nøje og fra flere synsvinkler, før man godtager det, men der er heller ingen mening i at forkaste statistikken og dens metoder i sig selv. Man fortryder jo ikke at man har lært at læse fordi nogle forfattere bruger sproget til at tilhylle sandheder og dølge sammenhænge i stedet for at stille dem blot for læseren. Der kan lyves uden statistik så vel som med.
Stjernebøgerne, bind ??
Omslaget viser en graf med to kurver på og en mand, der fejer nogle tal ind under gulvtæppet
Indskannet omslag - N650U - 150 dpi
Oversat fra amerikansk "How to Lie with Statistics" af Jens Munk-Madsen
Illustreret af Irving Geis
Side 51: Man blæser på den gode gamle regel, at en forskel kun er en forskel, hvis den gør en forskel.
Side 70: Hvis De vil være lige så sludrevorn som artiklens forfatter, kan De bevise at trafikrisikoen er større i klart vejr end i tåge, for der sker færre ulykker i tåge. Det er nemlig kun tåget vejr en mindre del af årets dage. Ikke desto mindre kan tåge være temmelig risikabel for trafikanterne.
Side 79: To ure går nøjagtig ens. Når uret a viser hel, slår b. Får uret a uret b til at slå?
Side 92: Så længe det er regelen at de fejl der forekommer, går til samme side hver gang, er det temmelig søgt at forklare dem som bommerter eller uheld.
Side 96: Dette ejendommelige stykke statistik overdriver på to måder. For det første går man ud fra det aritmetiske gennemsnit i stedet for midttallet, det har vi tidligere været inde på. For det andet bilder man sig tilsyneladende ind at en families indtægt bestemmes af familiemedlemmernes antal. Jeg har selv fire børn, og ville være særdeles tilfreds hvis det skulle blive indrettet sådan, men det tør jeg ikke håbe på. Familier med fire medlemmer er ingenlunde generelt bedre stillet end familier på to.
Side 96: .. det tjener til påmindelse om hvor lidt tiltro et gennemsnitstal er værd, når beregningsmåden ikke er oplyst.
Side 98: En procentberegning på grundlag af et lille antal er næsten altid vildledende. Antallet i sig selv er mere oplysende. Og når procentangivelsen har decimaler med, er det ikke bare pjattet, det nærmer sig fup.
Side 103: Hvis man ganske frimodigt lægger procentuelle stigninger sammen, ender man i det rene vanvid. Tænk Dem at De køber tyve forskellige varer i dag, og efter et års forløb viser det sig at hver af varerne er steget fem procent i pris. Fem procent lagt sammen tyve gange er 100 procent; prisniveauet er altså steget til det dobbelte på et år. Sikke noget sludder.
Side 106: Det kan undertiden være en hel bedrift at påvise at et kontroversielt emne kan anskues anderledes end i sort og hvidt.
Side 109: Konklusionen må blive at der er al mulig grund til at granske statistisk materiale i bøger, blade, og annoncer nøje og fra flere synsvinkler, før man godtager det, men der er heller ingen mening i at forkaste statistikken og dens metoder i sig selv. Man fortryder jo ikke at man har lært at læse fordi nogle forfattere bruger sproget til at tilhylle sandheder og dølge sammenhænge i stedet for at stille dem blot for læseren. Der kan lyves uden statistik så vel som med.
Stjernebøgerne, bind ??
Similar in this library
Pages
126
DDC/MDS
| 519.5 |