What is the formal definition of the prior, likelihood, and posterior in the context of the librarian vs. farmer problem?
Conditions
- Hypothesis H: 'Steve is a librarian'
- Evidence E: 'Fits the description'
- Population assumption: 10 librarians, 200 farmers
- Likelihood assumption: 40% for librarians, 10% for farmers
Reasoning, step by step
- Define the hypothesis H and evidence E.
- Calculate the prior from the base rates ().
- Identify the likelihood P(E|H) from the descriptive match rate (0.4).
- Compute the posterior P(H|E) using Bayes' theorem or the sample grid.
- Interpret the posterior as the updated belief.
Example
"Let H be the hypothesis 'Steve is a librarian' and E be the evidence 'fits the description'. The term is the prior probability. The term P(E|H) = 0.4 is the likelihood—how probable the evidence is if the hypothesis is true... This result, P(H|E), is called the posterior—the updated belief after incorporating new data."
Common misconceptions
- Confusing the prior with the posterior.
- Thinking the likelihood is the probability of the hypothesis given the evidence.
- Assuming the prior reflects real-world statistics rather than the specific illustrative setup.
Watch the explanation
Connected concepts
Explore next
Related questions
Starting from the equality (B|A) = (A|B), you can solve for either conditional probability by dividing by the corresponding marginal probability. Dividing both sides by isolates P(A|B), giving P(A|B) = (B|A)/P(B).
Conditions: Both A and B have positive probability for the ordinary conditionals used here.; The algebraic rearrangement requires the denominators and to be nonzero.
The conjunction fallacy manifests when individuals judge the probability of a combined event (being a bank teller AND active in the feminist movement) as higher than the probability of one of its constituent parts (being a bank teller). Mathematically, the second event is a subset of the first, so .
Conditions: Comparing the probability of a subset event against its superset event; Events are defined such that one is contained within the other
The event 'A and B' is logically identical to 'B and A', so any valid decomposition of its probability must agree. This symmetry forces the equality of the two product formulas: (B|A) = (A|B).
Conditions: The argument uses commutativity of logical conjunction for events.; Both A and B have positive probability for the ordinary conditionals used here.
The formula P(A and B) = is crossed out because it is only valid for independent events. The video emphasizes the general multiplication rule, P(A and B) = (B|A), which works for both independent and dependent events.
Conditions: The events A and B may be dependent.; The general multiplication rule P(A and B) = (B|A) applies regardless of independence (assuming ).
The calculation assumes a population of 210 people: 10 librarians and 200 farmers. With 40% of librarians fitting the description (yielding 4 matching librarians) and 10% of farmers fitting it (yielding 20 matching farmers), there are 24 total matches.
Conditions: Population consists only of librarians and farmers with a 1:20 ratio; Likelihoods are stipulated as 40% for librarians and 10% for farmers; Conditioning requires
Answers are generated from source material and independently checked. Consult the original video or creator if something is unclear.