Braineos
Sign in
← Back to decks
Sign in to save your progress, vote, and build your own decks.Sign in

Stats that fool you

The everyday mistakes behind misleading numbers and charts: correlation, base rates, survivorship bias, cherry-picking and more.

16 cards·by braineos-team
datacritical-thinkingstatistics
Study this deckPlay Tower Defense

Where this fits

data

No other decks with this tag yet.

critical-thinking

No other decks with this tag yet.

statistics
AP Statistics Chapter 8
0 cards
Correlation vs causation
Two things moving together doesn't show that one causes the other.
Base rate
How common something is to begin with. Ignore it and rare-event results look more certain than they are.
Survivorship bias
Drawing conclusions only from the cases that made it through, while the failures are out of sight.
Cherry-picking
Choosing only the data, or the time period, that supports a point.
Sample size
How many observations a result rests on. Small samples swing a lot by chance.
Regression to the mean
An extreme result tends to be followed by a more ordinary one, even if nothing changed.
Relative vs absolute risk
"Doubles the risk" can mean going from 1 in 10,000 to 2 in 10,000.
Truncated axis
A chart whose scale doesn't start at zero, which makes small differences look big.
Confounder
A hidden factor that affects both of the things being compared.
Selection bias
When the people or data studied differ from the group the conclusion is about.
Mean vs median
A few extreme values pull the mean. The median is the middle value and isn't pulled.
Ice cream sales and drownings both rise in summer. What explains the link?
  1. Ice cream causes drowning
  2. ✓ A confounder: hot weather
  3. Regression to the mean
  4. A small sample
A test is 99% accurate for a condition 1 in 10,000 people have. You test positive. What matters most?
  1. The test's brand
  2. ✓ The base rate
  3. The time of day you took it
  4. The chart's colours
"Famous founders dropped out of college, so dropping out helps." Which mistake is this?
  1. ✓ Survivorship bias
  2. A truncated axis
  3. Absolute risk
  4. Regression to the mean
A team's worst month is followed by a better one after a new manager starts. What else could explain it?
  1. ✓ Regression to the mean
  2. The base rate
  3. The median
  4. A truncated axis
In an income chart, the mean is far above the median. What does that suggest?
  1. The data is wrong
  2. ✓ A few very high incomes pull the mean up
  3. Most incomes are above the mean
  4. The sample is too large