PSYCH 240-Lecture 21

Probabilistic.
Conclusions that are likely, but not necessarily true.
Discussed in “Judgment and Decision Making” lecture

If the disease is present, the test will be positive 80% of the time.

If the disease is absent, the test will be positive only 10% of the time.

BUT -- If only 1% (10 out of 1000) people actually have the disease:

Current evidence
Hit Rate
False Alarm Rate

Base rate

Takes 3 things into account:
The BASE RATE - P(H)
The HIT rate - P(E/H)
The FALSE ALARM rate - P(E/not H)

Bayes’ theorem simply describes how to do the computation on these 3 pieces of information

A problem is base rate neglect: people often ignore the base rate and pay too much attention to the evidence.

People often rely too much on the current evidence and ignore the base rate.
People put more emphasis on similarity than on base rate
2 blue and 1 red is more similar to 20 blue and 10 red (bag B)

Theory: People rely on heuristics in judgments under uncertainty
rules of thumb are quick, require fewer cognitive resources, are “good enough” for everyday reasoning.
Lab studies exploit the weaknesses of the heuristics

Judge whether A comes from class B by relying on the similarity of A to B.

e.g., judge whether your patient is schizophrenic by relying on how similar he or she is to schizophrenics.

Problem: Focuses too much on current evidence - ignores base rates

People ignore base rate because of biases and expectations

neutral description

When given no other information, subjects can use base rate information to judge probability
Part 1

When other information is available (even when it is uninformative!), base rate information is ignored and people use representativeness (similarity) as the basis for probability judgment

Normative reasoning is cognitively demanding (WM!)
Heuristics are usually “good enough”