Norms, Confirmation Bias, Quick Judgements

Schema

A set of concepts linked in the semantic network — things that naturally go together in your mind

Schema-consistent items make it easier to access other related items

Raises accessibility/familiarity

• Ex. thinking of "toast" might bring "eggs" or "bacon" to mind

Schemas represent:

• Common patterns of experience

• Rules or sequences of behavior

Schemas x culture

Schemas can vary by culture

🍳 English breakfast → toast, eggs, bacon, beans, grilled tomato

🍚 Japanese breakfast → rice, fish, soup

Everyone’s schema is shaped by culture, experience, and context!

T or F: Schemas help make sense of ambiguity

Schemas automatically activate in your mind, helping you make sense of ambiguous situations

• They activate so fast you don’t even realize it's happening!

Schemas provide context, fill in gaps, and guide interpretation

Ex. McGurk Effect👄

• When you hear one sound (like "ba") but see lips mouthing another (like "ga") your brain is faced with ambiguous input

• Speech schema uses visual cues to reinterpret what you hear

Schemas x unlikely events

System 1 doesn’t care about how likely something is — it cares about pattern recognition

Even odd or fictional schemas (ex. zombie apocalypse, shark attacks) can be triggered when cues match

Ex. You're paddling on Lake Ontario and see ominous fins...

• You rationally know that there are no sharks in the lake

• But your System 1 activates the "shark attack" schema: SHARK! DANGER! PADDLE FASTER!

• The schema is automatically activated, even though the odds of it actually being a shark is low...

System 1 x Surprise

(1) Schema-inconsistent info -> (2) SURPRISE! -> (3) Schema updates

(2) Suprises happen when our system 1 detects something that does not fit with our schemas

• System 1 does not like surprises…

• Suprises = failed predictions

 

(3) We update our schemas so that we will not be surprised again

T or F: Schema-updating is hypersensitive

TRUE! And this is problematic….

 

It only takes one instance for an unexpected, schema-inconsistent thing to become expected….

We are liable to incorporating surprising info into a schema too quickly

Examples of system 1 processes 

Schema activation

→ Instantly fills in mental patterns when cues are detected

Schema updating

→ Adjusts mental models after new experiences (to avoid surprise)

McGurk Effect

→ Combines sight + sound into a coherent perception, even if wrong

Schema-consistent false memories

→ Fills in memory gaps with what “should” be there based on expectations

Causal intuitions

→ Automatically searches for simple causes and story-like explanation

System 1 x Causal intuitions (seeing patterns that are not there…) 

Causal intuitions = system 1 has a thematic tendency to search for causes and see simple stories

• This can be a problem because not everything means something…

System 1 often:

• Sees causality where there is none

• Forces ambiguity into neat schemas

• Creates “meaning” out of randomness

Two triangle + one circle (abstract video study)

What's going on in this clip??? (no audio, two triangles, one circle, going inside and out of a rectangle)

• Common participant interpretation = bullying, violence

• Even though this clip simply shows shapes moving around…

 

Study designed to demonstrate how people try to find meaning in simple clips-- we derive intentionality, narratives, emotions 

• Illustrates how humans are very good at constructing stories! (S1 automatically makes a story)

• We search for stories and schemas to make sense of what is happening

T or F: System 1 is biased to believe

Study details:

Participants first given various nonsense statements

• Ex. "A dinca is a flame = True"

• Ex. "A yerlow is a horse = False"

 

Half of the participants had to do this task while under cognitive load (had to remember a random 6 digit number sequence)

• This second task involves rehearsal which takes up system 2 cognitive processing!

 

Participants then asked whether the statements they saw earlier were true of false?

 

Findings:

Cognitive load had selective effects on the kinds of mistakes people made

Participants who were under cognitive load were more likely to think originally false statements were true

• Cognitive load means we can only use S1 processing, meaning we are way more likely to believe everything we see (which explains why they thought false statements were true)

Rule discovery

When we have a hypothesis, we try to find evidence that confirms the hypothesis rather than evidence that disconfirms it

Works well when your hypothesis is correct

BUT problematic when it's wrong, because you don’t test alternatives…

Our brains are better at saying "Does this fit?" than "Does this contradict?"…

Wason card selection task

Task is a way of studying confirmation bias

Rule: If there is a vowel on one side of the card, then there is an even number on the other. You must choose which cards you to turn over to check if the rule is being followed!

Cards shown: A K 4 7

What most participants choose: A 4

• But this is not the right approach…

• This is confirmation bias

Correct cards to turn over (*only cards that could break the rules):

A – to check if there's an even number (must follow rule)

• If not, the rule is broken!

7 – to check if there's a vowel (would break the rule)

• If there is, the rule is broken!

Why note 4 or K:

4: Rule doesn’t say the opposite— "if even, then vowel"

K: Rule only applies to vowels

Key point:

You need to avoid confirmation bias — look for cards that could disprove the rule, not just confirm it

Wason card selection task (drinking version)

Rule: If drinking alcohol, you have to be 19 years old

Who do you have to check in order to know whether everyone at the party is following the law?

- Check the 16 year old + the person with beer

- You wouldn't check the 25 year old because they can drink anything, and you wouldn't check the soda because it can be drank by anyone of any age…

Turning a problem into one that you already have accurate schemas for is helpful for decision making!

Individual differences in jumping to conclusions (JTC)

System 1 jumps to conclusion and seek confirming evidence

• But there are individual differences in JTC tendencies

 

Study details: Jimmy is fishing in 1 of 2 lakes:

• Lake Greyfish: 80% grey fish, 20% red fish

• Lake Redfish: 80% red fish, 20% grey fish

 

The first fish Jimmy catches is red. Do you think you know which lake he is fishing in, or would you like to see the colour of the next fish he caught? (Repeats 5x)

 

Findings:

~30% of people will make a guess about the lake after seeing just 1 fish

Low #  of fish requested = higher JTC bias

High # of fish requested = lower JTC bias

• More accurate judgement

• More evidence-informed decision making

Halo effects x jumping to conclusion

Tendency to think that positive traits co-occur with other positive traits (or negative with negative)

System 1 schema updating is very sensitive (can be done on the basis on little info)

  

Study details:

Participants given character description for either Koji or Iris

• Koji = intelligent, industrious, impulsive, critical, stubborn, envious

• Iris = envious, stubborn, critical, impulsive, industrious, intelligent

Participants then asked to rate how likeable Koji or Iris are

 

Findings:

Participants tended to think . . .

• Koji = quite likeable

• Iris = quite dislikeable

 

Even though they have the same traits, why do they get different ratings…

• Halo effect -> after participants saw the first trait (intelligent vs. envious), they tend to fill in the blanks using that info

• Primacy effect -> paying more attention to the first thing they see (intelligent vs. envious) compared to the last (envious vs. intelligent) 

• Priming + conceptual accessibility -> a lot of the traits in the middle of the list are ambiguous (ex. it is unclear whether being "critical" is a good or bad thing), thus the first trait in the list ends up colouring the rest of the traits in the list

  • Intelligence + critical = good thing, they may be a critical thinker

  • Envious + critical = bad thing, they may be a jealous person

 

Thin slicing (the power of first impressions)

Study details:

Researchers sit in on different psyc prof's first day of lecture and record the whole lecture

• The lecture recording is then cut down to simply the first 10 seconds of the class (AKA the first impression of the prof)

 

Uni students in the lab are shown the 10 second clips + asked to rate the profs (likeable? Competent?)

• These ratings are compared to the prof's final ratings at the end of the semester

 

Research Q:

Does first impression predict teaching ratings at the end of the year?

 

Findings:

YES! First impressions DO matter!

The more likeable and competent profs seemed in first 10 seconds = higher prof ratings at the end of the year

WYSIATI ("what you see is all there is")

System 1 only considers what it sees, and thinks that what it sees is true and sufficient

• The info that we do not see / info that is not immediately available to you may as well not exist in terms of your system 1 judgement process…

T or F: System 1 jumps to conclusions rather than waiting for more info

TRUE 

T or F: System 1 creates schemas that are insensitive to likelihood

TRUE

T or F: System 1 uses the first thought that comes to mind

TRUE

T or F: System 1 trusts feelings associated with thinking (fluency)

TRUE

T or F: System 1 is biased to believe, finds stories

TRUE 

T or F: We prefer coherence over completeness

TRUE!

It feels good to have limited info (and feels bad to have all of the info)

• Ample info that is inconsistent = DISLIKED :(

• Limited info that is coherent = LIKED :)

 

Study details:

Participants provided with an ambiguous legal scenario

Participants split into three groups:

1. Only hear the defendant's arguments (one sided)

2. Only hear the prosecution's arguments (one sided)

3. Hears all the arguments (fair)

*Participants are told whether they are in a one-sided group

 

Participants asked (1) how guilty is the defendant? (2) how confident are you in your judgement?

 

Findings:

Participants who only heard defense's argument thought the defendant was less guilty

• If you only hear info about the defendant, you are more likely to believe the defendant

 

Participants who heard only one side = very confident

• Less info leads to more confident judgements

Participants who heard all arguments = less confident

• More info leads to less confident judgements

 

*From a rational standpoint, these results do not make sense… you would think that more info gives you a better basis for making decisions / being confident…

Heuristic ingredients (basic assessments)

System 1 makes basic assessments with little effort all the time!

• How do I feel right now?

• Is X good or bad?

• How are things going?

• Do things feel normal?

• Should I approach or avoid X?

• Is person X friend or foe?

What are the heuristics ingredients?

Basic assessments

Intensity matching

Mental shotgun

Heuristics ingredients (intensity matching) 

The brain’s ability to match intensity across different domains (even if those domains are unrelated) 

• System 1 can intuitively say: “This thing feels as extreme as that thing — so they must go together”

Heuristics ingredients (mental shotgun)

When you have to make a complex JDM, system 1 automatically supplies lots of unnecessary information

This distract actually relevant information!