Reasoning and Decision Making I

Question 1

people don’t naturally understand stats study

Answer 1

Teigen (1983)

4 suspects in murder scenario, mean probability judgment was 130%

5 suspects: 159% (prior probs didn’t change)

2 suspects: 100%, more intutitive

Question 2

how to overcome errors of intuitive judgement

Answer 2

Robinson
and Hastie (1985) 

participants read murder scenario, effect avoided with training

Question 3

availability heuristic shit evidence

Answer 3

Lichtenstein (1978) overestimate rare causes of death, underestimate common ones (BUT about environment, not direct judgement, could also be central tendency judgement)

Question 4

availability heuristic actual evidence

Answer 4

T+K (1989) STUDY 1

19 famous men, 20 less famous women (and vice versa)

80 out of 99 participants judged that this list had more men overall, because they remembered more (more were available)

T+K (1989) STUDY 2

participants estimated that a novel would contain less –n- ending words than -ing ending words

conjunction fallacy (prob can’t be less!). More -ing words available.

Question 5

availability heuristic

Answer 5

judgments are based on the ease with which relevant instances come to mind

Question 6

representativeness heuristic

Answer 6

judgments are based on the extent to which an outcome (or item) is representative of the process or category in question

Question 7

representativeness heuristic evidence

Answer 7

T+K (1973) STUDY 1

base rates: 30 engineers, 70 lawyers (or vice versa)

people judged probability of jack being a lawyer as largely the same when drawn from majority-engineers population as when majority-lawyers population

T+K (1973) STUDY 2

• prediction group: personality sketch similar to engineer, predicted likelihood of 9 subjects
• base-rate group: not given sketch, but given national base rates of 9 subjects
• representativeness group: asked about how rep he fitted with stereotypes of 9 subjects

PREDICTION group highly correlated with representativeness judgements, negatively correlated with base-rate

Question 8

inverse fallacy

Answer 8

probability (A | B) =/= (B | A)

probability (cancer | positive) =/= (positive | cancer)

Question 9

inverse fallacy study

Answer 9

Eddy (1982)

95 out of 100 physicians demonstrated the inverse fallacy

Question 10

issues with representativeness heuristic

Answer 10

1. base rate neglect not so common (Koehler, 1994)
2. representativeness not portrayed properly in those 2 studies
3. base rate neglect/inverse fallacy may not be heuristics (Gigerenzer)

Question 11

Gigerenzer study

Answer 11

“natural sampling” - we have evolved to keep track of preg over time

STUDY - reword physicians study, correct answers increased from 8% to 46%

Question 12

Gambler’s fallacy 2 studies

Answer 12

1. Croson and Sundali (2005)

Nevada casino red/black or odd/even

run-increased -> more likely to bet on opposite colour. streak of 6 - 85% said that the streak would end

2. Matthews (2013)

coin toss sequences, heads judges as less probably after streaks of 5 heads

Question 13

Hot hand studies

Answer 13

1. Gilovich et al. (1985) found that probability of scoring was independent of run, but fans believed otherwise
2. Ayton and
   Fischer (2004) roullette game, people’s confidence in their predictions increased after a run

Question 14

NO gamblers fallacy and hot hand cannot be explained by representative argument + study

Answer 14

Ayton and Fischer (2004)

• random outcomes (e.g. coin toss) are sampling without replacement - GF is overgeneralisation
• intentional human performance (e.g. learning to play basketball) = positive recency, we actually get better every time

STUDY

• sequences with different streak lengths
• participants were more likely to attribute long sequence to intentional human performance (like basketball) than to a random mechanical process like coin flipping.