Inductive logic: Common Errors When Estimating Prob

Question 0

Gambler’s fallacy

Answer 0

Believing probs are which are independent are dependent

Question 1

The conjunction fallacy

Answer 1

Prob 2 propositions A + B both being true = prob (A is true) x prob (b=true)
– A + B is less likely to be true than either A or B on their own

Question 2

Estimating coincidences

Answer 2

People are prone to underestimate coincidences, especially when number of cases are very large

Question 3

Generalising from a small sample

Answer 3

Risky (law of large numbers)

Sample size greater than 400

Question 4

Generalising from unrepresentative sample

Answer 4

Obtained only through randomisation

Question 5

Self-selecting polls

Answer 5

Only volunteers take part in it= not representative of pop as whole

Question 6

P conditional on Q not the same as Q conditional on P

Answer 6

EG.
Pick an exeter at student= 4/5 chance British
Pick British person = not 4/5 chance at exeter

Question 7

Base rate fallacy

Answer 7

Estimating prob of (P given evidence E) = have to take into account prob of P on its own
– if very low, unlikely that E will make much diff

Question 8

Confusing correlation and causation

Answer 8

P + Q = correlated doesn’t necessarily mean P causes Q

– might be a 3rd thing

Question 9

Confirmation bias

Answer 9

We have a hypotheses/fixed belied, ten to focus on evidence that confirms it and ignore evidence that goes against