Post-Midterm II: April 8-April 11 Flashcards
general note for understanding when children “learn X”
children VARY A LOT in when they develop certain capacities
age ranges should be viewed like a measure of CENTRAL TENDENCY (ie. mean, median) rather than a rule
nonverbal recognition: how to measure recognition of social categories?
preverbal infants can’t tell us if they recognize something
can INFER NOVELTY from EYE GAZE DURATION
habituation paradigm
habituation paradigm: looking time can assess both…
- novelty
a) recognition of novelty, difference or change
- preference
habituation paradigm: recognition of novelty, difference or change
when a pre-existing stimulus has been COMPLETELY ENCODED
habituation paradigm: preference
when a pre-existing stimulus HASN’T BEEN FULLY ENCODED
what do infants do when a pre-existing stimulus hasn’t been fully encoded?
they show a PREFERENCE for it
novelty example: habituation paradigm - sample
5-6 month olds
novelty example: habituation paradigm - design
- habituate to a photo
- see a new photo of someone who is SIMILAR or DIFFERENT in GENDER or AGE
- assess looking time
novelty example: habituation paradigm - DV
looking time at new photo
habituation = complete encoding, so recognition
preference example: habituation paradigm - sample
white newborns and white 3 month olds
preference example: habituation paradigm - design
- saw images of people from different races
NO HABITUATION
- assess looking time
preference example: habituation paradigm - DV
looking time
novelty example: habituation paradigm - RESULTS
LESS looking time: for similar gender and age
MORE looking time: for different gender and age
preference example: habituation paradigm - since there’s no habituation in this manipulation…
looking time reflects PREFERENCE
preference example: habituation paradigm - RESULTS
NEWBORNS:
a) no race-based difference in looking time
3 MONTH OLDS:
a) more looking time at WHITE FACES
preference example: habituation paradigm - MAIN POINT
nonverbal preferences based on race develop with experience
because newborns showed no diffs in looking time
but 3 month olds looked more at racial ingroup (preference)
2 rules for when awareness of social categories develop
- VISIBILITY of social category
- PRIMACY of social category in EVERYDAY life
implications of visibility of a social category for what social categories kids perceive first
younger children are most aware of physically visible categories
like RACE and GENDER
when does awareness of gender develop?
2-3 years old
children latch onto gender roles pretty strongly
when does awareness of race/ethnicity develop?
5 years and older
order of races that White children in the US can distinguish
Black people first
Asian, Latinx and Native American people later
but there’s lots of variation
do children and adults determine race in the same way?
no, they do this differently
children vs adults determining race study SAMPLE
adults and 4-9 year olds
in Northeast US
81% White
11% Black
8% other
children vs adults determining race study DESIGN
participants saw White and Black faces
had to label them as:
- “White or European American”
- “Black or African American”
faces differed in:
- skin colour
- facial features
children vs adults determining race study - how did the face stimuli differ?
- skin colour
- facial features (physiognomy)
there were more and less Afrocentric faces
children vs adults determining race study - participants had to label the faces as either…
- “White or European American”
- “Black or African American”
children vs adults determining race study - main point
adults rely on a combination of skin colour and facial features
children rely primarily on skin colour
what do children rely on when determining race?
skin colour
what do adults rely on when determining race?
skin colour
facial features
study: do kids think about race as essential? what did this study look at?
looked at the development of race as an IMMUTABLE or ESSENTIAL feature of a person
study: do kids think about race as essential? SAMPLE
5-6 year olds
9-10 year olds
adults
study: do kids think about race as essential? SETUP
5-6 year olds, 8-9 year olds and adults viewed images of:
White and Black kids and adults
that were expressing either a HAPPY or ANGRY expression
on each trial, participants saw a photo of one child who was making either a HAPPY or ANGRY expression and then:
a) one SAME-RACE adult making a DIFF expression
b) one OTHER-RACE adult making a SAME expression
participant had to answer “when the child grows up, which one will he be?”
study: do kids think about race as essential? PARTICIPANTS HAD TO ANSWER WHAT QUESTION?
“when the child grows up, which one will he be?”
study: do kids think about race as essential? RESULTS
9-10 year old White kids and adults viewed RACE AS MORE IMPORTANT than emotion when making predictions
5-6 year old White kids USED RACE MUCH LESS compared to White 9-10 year olds
BUT MINORITY 5-6 year olds were ALSO MORE LIKELY TO USE RACE
study: do kids think about race as essential? what groups were more likely to use race when making predictions?
9-10 year old White kids
White adults
racial minority 5-6 year olds
study: do kids think about race as essential? SPECIAL FINDING IN RACIAL MINORITY 5-6 YEAR OLDS
racial minority 5-6 year olds (unlike White 5-6 year olds) were more likely to use RACE than emotion when making predictions
as kids become more aware of race as an aspect of one’s social identity…
they also become aware of the cultural NORM of HESITANCY to DISCUSS RACE
“colourblindness”
children may learn to avoid discussing race even in situations where the situation might demand it
how far does “colourblind” norm go in kids?
children may learn to avoid discussing race even in situations where the situation might demand it
“guess who” study investigated what?
how far children would go in avoiding discussing race by using a modified version of the game “guess who?”
“guess who” study examines race as a _____ issue
sensitive
“guess who” study setup
show three photos of people
ask “which one is Thomas?”
participant must ask very specific questions to figure out who Thomas is
will kids hesitate to ask about race?
which age group will arrive at the answer quicker/ask the most efficient questions
“guess who” study conditions
- RACE-RELEVANT condition
a) pictures include people of diff races - RACE-NEUTRAL condition
a) pictures are of people of the same race
b) but have a sticker at the bottom of each photo - either brown or light beige
will kids use this sticker over over referencing race?
“guess who” study results
key difference when testing 8-9 year olds versus 10-11 year olds
- 8-9 year olds asked roughly the same amount of questions across both conditions
- 10-11 year olds needed MORE QUESTIONS to get to the right answer in the RACE-RELEVANT condition
SO YOUNGER KIDS DID BETTER ON THE TASK WHEN RACE WAS INTRODUCED
2 important factors for development of racial prejudice
- are you a member of a MAJORITY or MINORITY group?
- how much CONTACT do you have with members of minority groups?
“guess who” study takeaway
8-9 year olds aren’t affected highly by social norms about avoiding discussing race
10-11 year olds are though, so they take longer to accomplish “guess who” task
CHART: relative prejudice levels depending on group membership
lowest prejudice level:
MINORITY GROUP
middle prejudice level:
MAJORITY GROUP WITH HIGH OPPORTUNITY FOR CONTACT
highest prejudice level:
MAJORITY GROUP WITH LOW OPPORTUNITY FOR CONTACT
development of prejudice depends on…
majority/minority group status
amount of contact
CHART: changes in explicit racial prejudice as kids age
WHITE KIDS:
a) at age 5, like White people more and stay VERY CONSISTENT as they age
b) explicit liking of minorities increases with age
implicit attitudes across development STUDY METHOD
IAT
modified to work with kids
used pics of Black and White faces
and smiley/frowny faces instead of “good”/”bad”
implicit attitudes across development STUDY SAMPLE
N = 79
mostly White participants in Boston
27 kindergartners
30 fifth graders
22 adults
implicit attitudes across development STUDY RESULTS
- SELF-REPORTED preference for White over Black kids LOWERS AS AGE INCREASES
a) highest in 6 year olds, then lowers in 10 year olds and lowers again in adults
- IMPLICIT Pro-White Attitudes STAY STEADY ACROSS AGES
a) are the same for 6 and 10 year olds and adults
differences between implicit and explicit attitudes: why does EXPLICIT PREJUDICE DECLINE?
- learn social and cultural NORMS about race
- internalize MORAL lessons about EQUALITY & FAIRNESS
implicit attitudes across development TAKEAWAY
self reported favouring of Whites over Blacks decreases with age
implicit preference for Whites stays stable across ages
REFLECTS POWER OF SOCIAL NORMS
differences between implicit and explicit attitudes: reasons for STABILITY IN IMPLICIT ATTITUDES
- stability of ATTITUDE
- stability of CULTURAL MESSAGES
- implicit prejudice increases, but adults get better at controlling them
a) two things in opposing directions at work
b) an increase and a decrease
implicit and explicit race attitudes among children in Cameroon STUDY SETUP
investigated implicit and explicit race attitudes towards Black, White and Chinese people
among children growing up in Cameroon
- adapted IAT for implicit attitudes
- preferences for own-race vs other-race people in a variety of scenarios
implicit and explicit race attitudes among children in Cameroon STUDY SAMPLE
30 participants from following age ranges:
3-6
6-9
9-12
12-15
15-18
18-30
implicit and explicit race attitudes among children in Cameroon STUDY - sample scenario
sample scenario for testing explicit attitudes
“this summer your mother will take you to a swimming class. you can choose one person to coach you to swim. which one would you like to choose?”
implicit and explicit race attitudes among children in Cameroon STUDY RESULTS
results revealed DIFFERENCE ACROSS THE LIFESPAN for how implicit and explicit attitudes change over time
- IMPLICIT:
a) pre-age 5/6, show ingroup preference
b) pro-White or pro-Chinese preference emerges after 6
- EXPLICIT:
a) stay stable across ages
b) favour ingroup
implicit and explicit race attitudes among children in Cameroon STUDY - WHAT HAPPENS AT 5/6?
children in Cameroon begin to favour the outgroup over the ingroup
study: does extended experience with other-race nannies predict racial bias in the preschool years?
in sample of one hundred 3-6 year olds in Singapore
researchers compared those with and without “other-race” nanny experience
children completed measures of explicit and implicit racial bias
study SAMPLE: other-race nannies
one hundred 3-6 year olds in Singapore
other-race nannies study: MEASURES THEY COMPLETED
- explicit racial bias
- implicit racial bias
other-race nannies study: RESULTS
more contact with other-race nannies was associated with LESS INGROUP RACIAL PREFERENCE in EXPLICIT ATTITUDES
LENGTH of contact with an other-race nanny WASN’T associated with the AMOUNT of ingroup racial preference in IMPLICIT attitudes
other-race nannies study: did the length of contact with other-race nanny correlate with the amount of ingroup racial preference in implicit attitudes?
no
so length of contact didn’t modulate implicit attitudes
whereas for explicit attitudes, MORE CONTACT with other-race nannies resulted in less ingroup racial preference
preference for one’s own gender develops when?
ages 3-4
recall: gender awareness develops at around 2-3
when do gender preferences decline?
around puberty
probably because of heterosexual attraction
girls as young as 6 years old were less likely than boys to…
report that members of their gender are “really, really smart”
ie. asked “which person is really, really smart” and shown a picture of a man and a woman
6 year old girls are less likely to think girls can be “really, really smart” - these beliefs in turn…
predicted LESS INTEREST in activities that were believed to be for “really, really smart” people
development of gender stereotypes: while 5 year olds thought…
5 year olds though their OWN GENDER was MORE LIKELY to be smart…
starting at AGE 6, both girl and boy participants though MEN WERE SMARTER
two things that happen at age 6
- minority groups begin to implicitly favour the outgroup
- girls and boys think men are smarter
this aligns with the beginning of formal schooling in most countries
development of gender stereotypes: the difference between 5 and 6 year olds carries over…
into interest in GAMES that were described as being for “really really smart people”
AGE 5:
a) girls are, if anything, MORE INTERESTED than boys
AGE 6:
a) BOYS are more interested than girls
“draw a scientist” task meta-analysis
meta-analysis covering 40 years, 78 studies, over 20 k participants
found that GENDER BIAS in the “draw a scientist” task is DECREASING OVER TIME
but NOT SUBSTANTIALLY
“draw a scientist” task: what happens for GIRLS as they AGE?
RATE of drawing a male scientists INCREASES with AGE
story book intervention to reduce children’s racial biases: STUDY SETUP
investigated whether children’s racial biases could be changed by reading a children’s story that featured a CROSS-RACE FRIENDSHIP
- manipulated whether the person reading the story to the child was Black (outgroup) or White (ingroup)
- also measured each child’s RECONCILIATION SKILLS
story book intervention to reduce children’s racial biases: STUDY SAMPLE
sample split between:
- kindergarteners (5-6 years old)
- second-graders (7-8 years old)
reconciliation skills
measured in the storybook intervention study
ability to accept whether others’ judgments (that differ from their own) are VALID
expectation that other people will differ from you in ways that are ok
story book intervention to reduce children’s racial biases: STUDY RESULTS
KINDERGARTENERS thought BOTH Black and White readers would have MORE POSITIVE attitudes towards White people
SECOND-GRADERS thought Black and White would PREFER their RACIAL INGROUP
^this was particularly true for those high in reconciliation skills
STORYBOOK INTERVENTION HAD NO EFFECT ON CHILDREN’S OWN RACIAL ATTITUDES
story book intervention to reduce children’s racial biases: did it affect children’s own racial attitudes?
no
story book intervention to reduce children’s racial biases: the results for the second-graders were particularly strong in which individuals?
in those high in reconciliation skills
(recall result: thought the reader would have a racial ingroup preference)
story book intervention to reduce children’s racial biases: one reason for its lack of success
children mostly assumed that the person reading the book largely SHARED THEIR OWN RACIAL PREJUDICES
young children may not possess the “cognitive structure to engage with an anti-bias perspective”
story book intervention to reduce children’s racial biases: what may young children not possess?
the “cognitive structure to engage with an anti-bias perspective”
the storybook intervention illuminated a HURDLE for effective interventions for kids…
must understand and appreciate what COGNITIVE SKILLS (ie. generalizing) are needed for an anti-bias message to be effective
kids didn’t generalize from the cross race friendship in the story to their own outgroup attitudes
Frances Aboud/Professor Emerita
expert on developmental psychology
specifically in terms of intergroup attitudes, prejudice and bias
currently working on applying these findings in middle and low-income countries
experience with both lab-based and field research on parenting interventions
social neuroscience: not just about…
localization
for example, categorizing racially ambiguous targets has been shown to lead to GREATER ACTIVITY in the dorsolateral anterior cingulate cortex
but why does this matter? knowing where something happens just for the sake of knowing that fact doesn’t help us in social psychology
neuro-scientific methods and timing
neuroscience can reveal the TIMING of various psychological PROCESSES
not possible using self-report or even ‘implicit’ measures like the IAT
the IAT is a very fast self-report, but it still requires some time for completing a behaviour response (ie. pressing computer key)
timing: neuroscience methods can bypass…
behavioural delays
reveal the TRUE TIME COURSE of various processes
neuroscience and timing: refresher from social categorization
White participants passively viewed images of male and female Black and White people
some categorized images based on GENDER, others based on RACE
ERPs revealed that:
a) race is processed within 100 ms
b) gender is processed within 200 ms
IAT behaviour (categorization decision) occurs at how many ms?
500 ms
FACE ONSET
100 ms: N100/social category detection
P200: goal-directed attention to category
200 ms
N200: category conflict & response selection
300 ms
400 ms
500 ms: BEHAVIOUR
without neuroscience, we wouldn’t be able to determine the things that happen before 500 ms
from ERP findings, what can we conclude about social categorization?
social categorization is FAST
and therefore likely AUTOMATIC
IMPLICATIONS of the ERP findings that social categorization is FAST/AUTOMATIC
larger implications for thinking about how such processes COULD or COULDN’T ever be CONSCIOUSLY CONTROLLED
social categorization may occur so quickly that finding a way to STOP the process is unrealistic
better to focus on how to LESSEN any potentially NEGATIVE IMPLICATIONS of automatic categorization
since social categorization may occur so quickly, we should focus our efforts where?
on how to lessen any negative implications of automatic categorization
instead of trying to stop it
social neuroscience can use the larger literature in COGNITIVE NEUROSCIENCE to show…
connections between various processes
this info can advance theory
can lead to new insights into how such processes operate
neurosynth: leveraging cognitive neuroscience literature
neurosynth gives you info about studies that have examined a SPECIFIC BRAIN AREA
type a brain area into neurosynth
comes up with TYPES OF TASKS that elicit activation in certain parts of brain
wide variety of studies that don’t have a ton in common, but that activate the same brain area
although two tasks may seem very different, if they activate the same part of the brain…
there must be some similarity across them
neurosynth uses what kind of data?
fMRI
connections between processes: N200
N200 has been associated with:
a) RESPONSE SELECTION
b) CONFLICT PROCESSES
because it originates in dorsal anterior cingulate cortex
typical finding of larger N200 response to ingroup targets in race categorization tasks may reflect RESPONSE CONFLICT associated with making an ingroup classification
N200 - typical finding of larger N200 response to ingroup targets in race categorization may reflect…
response conflict associated with making an ingroup classification
we arrived at this information through realizing that N200 has been associated in other work with RESPONSE SELECTION and CONFLICT PROCESSES because it originates in dorsal anterior cingulate cortex
just as intergroup neuroscience can show connections between different types of psychological processes, it can also be informative…
by taking the REVERSE approach
how can neuroscience tease apart processes that appear to be similar?
differences within processes: how the brain treats STEREOTYPES versus PREJUDICE STUDY setup
stereotypes and prejudice are distinct but related constructs
- on each trial, White participants saw TWO FACES that were either:
a) both Black
b) both White
c) one Black, one White
(analyses only focus on the White-Black trials)
- in some blocks, they made a STEREOTYPICAL JUDGMENT
a) “which person is more athletic?”
in other blocks, they made a PREJUDICIAL JUDGMENT
a) “which person would you want to have as a friend?”
- after forced-choice task in fMRI, participants then completed 2 IATs
a) measured ease with which Black and White faces could be paired with words related to ‘MENTAL’ (educated, smart, genius) versus words related to ‘PHYSICAL’ (athletic, agile)
differences within processes: how the brain treats STEREOTYPES versus PREJUDICE STUDY results
one brain region was consistently MORE ACTIVATED during FRIENDSHIP (prejudice) than trait (stereotype) trials
a different region was consistently MORE ACTIVATED during TRAIT than friendship trials
so diff brain areas are more activated when making stereotype versus prejudice judgments
differences within processes: how the brain treats STEREOTYPES versus PREJUDICE STUDY follow-up analyses
found that one brain region was DIFFERENTIALLY associated with the 2 IAT scores
depending on the JUDGMENT that participants were making
- during FRIENDSHIP trials, activation in the left temporal pole was MORE ASSOCIATED with EVALUATIVE IAT scores
(good/bad - prejudice) - during ATHLETIC trials, activation in left temporal pole was more associated with STEREOTYPE IAT scores
(mental-physical)
a final benefit of neuroscience approaches
can be used as way of RESOLVING competing PREDICTIONS/PERSPECTIVES
that wouldn’t be able to be resolve using other methods
racial paralysis
where people high in motivation to not appear prejudiced work hard to avoid cross-racial comparisons
racial paralysis study SETUP
task was similar to stereotype/prejudice IAT fMRI study, with ONE KEY DIFF:
participants were given option to indicates they had NO GUT FEELING
^ they could OPT OUT of making the judgment
racial paralysis study RESULTS
participants were MORE LIKELY to “opt out” of trials involving FACES OF DIFF RACES
particularly when making judgments related to STEREOTYPICAL traits
ie. intelligent, hardworking
racial paralysis study OPEN QUESTION
what drives this opt-out behaviour?
is it about lack of cross-race familiarity?
or is it more to do with efforts to regulate prejudice?
racial paralysis study - fMRI found greater activation where?
in the DLPFC
for CROSS-RACE over same-race trials
EVEN when participants “opted out”
BUT this was PARTICULARLY TRUE for when judgments were STEREOTYPE-RELEVANT (honest, intelligent, reliable) versus stereotype-irrelevant
racial paralysis study - greater DLPFC activation was particularly true when?
for cross-race trials
when they were stereotype-relevant (honest, intelligent, reliable)
racial paralysis study - the brain region that was more active during CROSS-RACE and STEREOTYPICAL judgments is implicated in…
- SELF-CONSCIOUS EMOTIONS
- REGULATION of BEHAVIOURS/JUDGMENTS governed by strong SOCIAL/MORAL NORMS
implications of these regions in cross-race decisions offers support for our account that the FEAR OF APPEARING BIASED evoked by such situations leads to…
CONFLICT, greater REFLECTION and a resulting tendency to OPT-OUT
racial paralysis study - implications of the activated brain regions
(activates areas involved in SELF-CONSCIOUS EMOTIONS and BEHAVIOUR/JUDGMENT REGULATION)
implies that regions used in cross-race decisions offers support for idea that FEAR of APPEARING BIASED leads to conflict, greater reflection and tendency to opt out
racial paralysis study - fear of appearing biased leads to…
a. conflict
b. greater reflection
c. tendency to opt out
what study looked at the more basic process by which people decide WHO IS and WHO ISN’T a group member?
study where researchers used task to DISENTANGLE whether people rely more on
a. similarity
b. group structure
when evaluating new people as potential group members
study about evaluating potential group members: SETUP
- participants first indicated their OWN BELIEF about a number of policy issues (ie. death penalty support)
- then they learned about policy beliefs for THREE OTHER TARGETS
- then they had to ALIGN with one of the targets by choosing to side with them on an unknown policy position
study about evaluating potential group members: participants had to align with one of the targets by choosing…
to side with them on an unknown policy position
essentially, had to think about which of the two would be the most like themselves when deciding on an unknown policy
study about evaluating potential group members: what affected people’s target choice?
across conditions, the LATENT STRUCTURE of people’s preferences made it MORE or LESS EASY to form a group with one of the targets
as the distractor (target C) becomes more similar to target B, preferences for target B increase
like C becomes a stepping stone to one of the other targets - can see a group forming
dyadic similarity
all that matters is similarity when choosing between A and B
so if A and B are equally far from you, probability of choosing either one is equal
latent structure
using the behaviour of others to infer a consensus or group structure
so if C is between you and B, leaving A off to the side, you’re more likely to align with B
study about evaluating potential group members: the study design can help TEASE APART…
whether DYADIC SIMILARITY or LATENT STRUCTURE is at work
do we only use similarity when making alignment decisions?
or do we take larger group structure/consensus into account?
study about evaluating potential group members: fMRI analyses found that…
greater use of this LATENT STRUCTURE approach was more strongly associated with activity in the RIGHT ANTERIOR INSULA
other studies have found this same brain region to be key to more GENERAL STRUCTURAL LEARNING tasks
^non social tasks! like in processing components of a sentence in a reading task
study about evaluating potential group members: results suggest that GENERALIZED GROUP CONCEPTS…
generalized group concepts rely on DOMAIN-GENERAL CIRCUITRY associated with:
a. latent structure learning
b. encoding of stimuli’s functional significance
these neuroscience findings speak to some core issues of the “cognitive perspective” on stereotypes and prejudice
seems like the way we think about people is highly similar to the way we think in general
study about evaluating potential group members: shows that the way we think about people…
is highly similar to the way we think in general
nothing special about social group decision making in terms of the neural mechanisms they’re recruiting
future directions for intergroup neuroscience
still a young field
research is increasing rapidly
has benefitted from new tech and measures
new developments in neuroscience tech are leading to new possibilities for WHAT can be studied from a neuroscientific perspective
2 new tech developments for intergroup neuroscience
- MOBILE MEASUREMENT
a. portable/less invasive ERPs
- INTER-BRAIN SYNCHRONY
a. tracking multiple people’s brains at the same time
b. are they engaging the same areas?
c. does common engagement result in higher harmony/better team performance?
d. all done in real time
inter-brain synchrony in teams predicts…
collective performance
Mina Cikara
prof at Harvard
expert on prejudice, particular focus on how neuroscience can inform intergroup processes
winner of 3 early career research awards for work on intergroup relations
author on 2020 paper investigating how learning related to social groups is connected to the anterior insula
social network analysis (SNA)
process of investigating social structures through the use of networks
how does SNA characterize individuals and relationships?
characterizes networked structures/individuals as NODES
relationships between nodes are TIES, EDGES or LINKS
node
entity in a network (person, group)
dot
edge
aka tie, link
connection between nodes
line
directed edge
edge that has an ORIENTATION
ie. an arrow indicating popularity
ie. indicates reciprocal versus non-reciprocal relationship
distance
smallest number of edges needed to connect two nodes
centrality
importance of a node in the network
ie. how many edges a node contains
ie. how many cross-group edges a node has
essentially, what amount of leverage does a single node have in a network?
advantages of SNA
- provides insight that people may not be able to self-report
ie. who is actually most popular versus who is perceived as most popular
- can identify popular nodes for interventions
limitations of SNA
- analyses are only as good as how much of the network you cover
- people belong to multiple networks simultaneously (work, family, school, clubs, sports etc) so effects in one network may or may not carry over
Paluck, Shepherd and Aronow: SNA anti-bullying intervention SETUP
used SNA to design and assess effectiveness of anti-bullying intervention
identified SOCIAL REFERENTS: kids in school network that had many connections to other kids
reasoned these kids would receive more attention and be looked to for info about group norms
social referents received intervention
throughout intervention, were encouraged to become PUBLIC FACE of opposition to these conflicts
notably, the intervention lacked educational/persuasive unit regarding adult-defined problems at school
social referents
Paluck anti-bullying intervention
kids in school network that had many connections to other kids
SNA anti-bullying intervention SAMPLE
56 middle schoolers
randomly assigned to CONTROL and INTERVENTION condition
within intervention condition schools, random sample was selected to receive receive intervention
SNA anti-bullying intervention INTERVENTION
meeting with research assistant every other week and discussing common conflict behaviours at school
student-led ideas on how to reduce the conflict
were encouraged to become the public face of opposition to these conflict
SNA anti-bullying intervention NOTABLE DETAIL
the intervention lacked an EDUCATIONAL or PERSUASIVE UNIT regarding adult-defined problems at school
SNA anti-bullying intervention RESULTS
schools in treatment condition SAW LESS CONFLICT than those in control
less disciplinary events for peer conflict
results were presented visually using SNA
NA anti-bullying intervention RESULTS WERE MORE EFFECTIVE WHEN…
intervention was more effective when a GREATER NUMBER of the ‘seed’ students in intervention condition were SOCIAL REFERENTS
as seed students GAIN CENTRALITY, get a larger reduction in peer conflict
Parkinson et al used SNA to do what?
to see how people encode social networks AUTOMATICALLY
an entire cohort of business school students were surveyed about their social network (who they liked to spend time with)
Parkinson et al SNA setup
- surveyed whole cohort of business school students about their social network
- subset was PASSIVELY SHOWN images of other members of social network while in fMRI
a. faces varied in degree to which they were DISTANT from participant in network (direct connection, friend of friend, friend of friend of friend)
Parkinson et al SNA results
distance in the SNA was consistently related to ACTIVATION IN 3 BRAIN AREAS
indicates that when encountering familiar individuals, humans may SPONTANEOUSLY retrieve knowledge of WHERE THEY’RE LOCATED relative to oneself
have a mental map of social spaces
distributional language analysis
reviewing large text bodies from a culture to identify what words are most likely to CO-OCCUR with one another
core assumption: MEANING of a word can be described by the words it co-occurs with
words occurring in similar contexts tend to have similar meanings
ie. “dog” is more similar to “cat” than to “banana”, because of the contexts they appear in
advantages of distributional language analysis
- insight that people may not be able to self-report
ie. what patterns exist in what we read
- allows for possible historical analyses for associations that may have existed before modern measures
ie. archival analysis
limitations of distributional language analysis
- needs a LOT of data (tens of millions of words)
- may be dependent on the type of text used (subtitles vs Reddit vs Wiki)
- can’t know whether text is SUPPORTING or REFLECTING or REFUTING certain associations
a. because simply tracks co-occurrence
ie. “men are better at work” = supporting
ie. “the culture believes men are better at work” - reflecting
ie. “it’s impossible that men are better at work” - refuting
across 25 languages, what did Lewis & Lupyan investigate?
the association between:
- the strength of GENDER-CAREER STEREOTYPE in a distributional language analysis
a. how closely man/career and woman/home co-occur versus the opposite pairing
- the strength of GENDER-CAREER IAT effect among participants speaking that language from Project Implicit
language may not only reflect pre-existing stereotypes…
it may also provide a distinct SOURCE of information for LEARNING ABOUT THEM
Lewis & Lupyan 25 languages study RESULTS
IMPLICIT (but not explicit) gender associations of participants in a country is correlated with gender associations embedded in the dominant language spoken in that country
since studies mapping attitudes onto distributional language analysis are…
correlational
it’s hard to determine causation
to what extend does the language we make/see CREATE implicit gender stereotypes?
to what extent do our implicit gender stereotypes CREATE changes in our language?
big data
social science approach
refers to an ARCHIVAL ANALYSIS that uses a dataset TOO LARGE TO CODE BY HAND
instead relies on AUTOMATED DATA COLLECTION and ANALYSIS
ie. web-scraping of tweets
advantages of big data approach
- same strengths as archival research, but with greater scope
- often measures impactful, real-world behaviour
- access to large amounts of data that make findings unlikely to be a fluke
limitations of big data approach
- same limitations as archival research but now with less ability to notice possible errors in data collection
- analysis dependent on data that’s available
- hard to get at causality
big data and medical records setup
big data approach applied to medical records: specifically TREATMENT ALGORITHMS
worked with large hospital to track treatment records of 50 k patients over 3 years
specifically investigated enrolment in “high risk care management program” which provided ADDITIONAL RESOURCES to patients with complex health needs
enrolment in program is primarily determined by ‘RISK ALGORITHM’ based on patient records
big data and medical records: what did enrolment in program depend on?
a Risk Algorithm
based on patient records
higher scores led to AUTOMATIC ENROLMENT in the program
big data and medical records RESULTS
even when Black and White patients had the SAME RISK SCORE determined by the algorithm…
Black patients had WORSE OBJECTIVE HEALTH MEASURES
ie. for a Black and White patient to earn the same risk score to gain eligibility, Black patients had 26% more chronic illnesses
big data and medical records: for Bl and Wh patient to earn the same risk score…
Black patients typically had 25% more chronic illnesses
big data and medical records: unlike most instances of research on algorithmic bias…
researchers had DIRECT ACCESS to how algorithm was working
it DOESN’T factor in patient race
yet Wh patients are given worse health scores than equally sick Bl patients - prioritizing their admittance to treatment program
big data and medical records: what did they find the algorithm was using?
HEALTH COSTS (how much patients spend to maintain their healthy) to predict overall health
but there’s a difference between “receiving health care” (health costs) and “needing health care” (objective health
big data and medical records: what introduced a racial disparity?
treating health costs as a proxy for health needs
Black patients generate less healthcare costs than White patients
because of a variety of SEC reasons
2 future directions and issues
- intersectionality
- algorithmic bias
intersectionality
the interconnected nature of social categorizations such as race, class, gender
regarded as creating overlapping and interdependent systems of discrimination or disadvantage
a theoretical approach based on such a premise
intersectionality: NY stop and frisk data setup
2006-2012
Black and White men (non-Hispanic)
64-76 inches (don’t differ in height)
photo ID only
1 073 536 valid targets
intersectionality: NY stop and frisk data results
5’4”: if you’re short, there’s a racial disparity in if you’re stopped and frisked (more Blacks)
5’10”: disparity increases
6’4”: disparity increases even more
intersectionality: NY stop and frisk data takeaway
if you’re Black in NYC, the taller you are the more negatively it affects you when it comes to stop and frisk
more likely to be stopped and frisked the taller you are
taller height is beneficial for white people, and harmful for black people
algorithmic bias
systematic and repeatable errors in a computer system that create unfair outcomes
like privileging one arbitrary group of users over others
coded bias movie
example of algorithmic bias
apartment building that’s being surveilled
speaks about obtrusive tech and AI influencing all kinds of decision making
feeds, ads, view of world is governed by AI
algorithms determine if you get into college, if you’re credit-worthy or not
Apple’s new credit card has been accused of sexist algorithms
Amazon hiring algorithm was found to be biased against women
quote about algorithmic biases
“the past dwells within our algorithms”
we train algorithms on biased input
ie. current tech employees are mostly men, so algorithms are biased towards selecting men
