Title How to Lie with Statistics
Author Darrell Huff
Year Published 1954
Kind of Book Statistics/ Rationality
How strongly I recommend it 9/10
My Impressions From the book: "This book is a sort of primer in ways to use statistics to deceive…honest men must learn them in self-defense." That pretty much sums it up. This is a classic. Short and sweet with funny illustrations to drive home each point.
Date Read Aug 2020
Practical Takeaways
Beware of statistics that are surprisingly precise (sim. to beware of even numbers)
Make sure the sample is representative of the whole (selection bias)
Allow yourself some degree of skepticism about the results as long as there is a possibility of bias somewhere
To determine if the sample is truly random: ask yourself "does every name or thing in the whole group have an equal chance to be in the sample?
Don't accept vague categories for stats ie. Ask which Americans this stat applies to
When you hear the word 'average' ask if it refers to 'mean, median, or mode'
When you see a statistic for average pay, first ask 'who is included?'
Put little funny cartoons on every couple of pages related to what you're talking about to engage and reward the reader. (they will read more seeing that there are little rewards sprinkled in)
Don't confuse 'normal' with 'desirable' (naturalistic fallacy)
Give a normal range in the statistic instead of a number when necessary (eg. Average IQ is x vs average Iq is between x-x)
Be careful not to include any highly emotional content without hastily saying whether you are for or against it
Make sure the graph you are reading starts at 0
Be aware of graphs that use very small units (y-axis) in order to make the findings appear more drastic than they are
Look with suspicion on bar graphs that change their widths or lengths while representing a single factor
Be aware of data represented as pictures or cartoons instead of graphs
Beware of images that are inaccurately drawn to represent data
If you can't prove what you want to prove, demonstrate something else and pretend that they are the same thing
Ask yourself if the data is relevant to what you're concerned with or only partially related
When looking at before or after pics study other variables between the pics like lighting, posture, hairstyle, etc.
Whenever you hear a statistic or a datapoint ask
Who says so?
How does he know?
What/who's missing from the sample?
Did someone change the subject? Ie. Is the statistic relevant to the question at hand?
Does it make sense?
Big Ideas
Statistics is as much an art as it is a science
Surprising Facts
The statistic average family size is 3.6, but this doesn't mean most families have 3.6 people. 3-4 person families make up 45% of the total. 35% are 1-2 and 20% have more the 4 people.
Colds last about a week no matter what you do (medicine only helps alleviate symptoms, but doesn't get rid of the cold sooner)
Unknown Terms
Literary Digest Error: A sampling bias toward the person with more money, more education, more information, better appearance, more conventional behavior, and more settled habits than the average of the population he is chose to represent. 2) When an interviewer is assigned to ask get the opinion of a person of a specific group (black or Texan or bearded) They are going to be more drawn to select people at the top of that group ie. Those most attractive, friendly, approachable etc. This leads to a sampling bias
Normal Distribution: If you draw a curve to represent the statistic, you get something shaped like a bell. Mean, median, and mode fall at the same point
Standard Error: The standard deviation of multiple means of a data set (eg. 7+ or- 2)
Standard Deviation: a measure of the amount of variation or dispersion of a set of values. A low number indicates that the values tend to be close to the mean of the set, while a high number indicates that the values are spread out over a wider range
Eye appeal: the quality of appealing to the eye; attractiveness
Yellow Journalism (yellow press): journalism and associated newspapers that present little or no legitimate, well-researched news while instead using eye-catching headlines for increased sales. Techniques may include exaggerations of news events, scandal-mongering, or sensationalism. The term yellow is used today as a pejorative to decry any journalism that treats news in an unprofessional or unethical fashion
Semi-attached figure: a situation in which one idea cannot be proven, so the author pulls the old bait-and-switch, stating a completely different idea and pretending it is the same thing. (eg. You can't prove that nostrum cures colds, buy you can publish that half an once of it kills 31,108 germs)