0:00

Welcome

0:04

to the Psych Podcast.

0:14

This is the podcast where

0:16

we talk about everything intro psych. I'm joined

0:18

as always by my friend and my colleague, Paul

0:21

Blum. Welcome, Paul. Hi, David. Good

0:24

to talk to you again. We're going to talk today

0:26

about social psychology. We're going

0:29

to continue a discussion from our previous

0:31

episode. And our topic is going to

0:33

be a topic which

0:36

is of interest to people both within and

0:38

outside the field. It's how we think about

0:40

human groups,

0:42

including groups based on

0:44

gender, on sex, on

0:46

ethnicity, on race, on age,

0:49

and how we stereotype the origins

0:51

of prejudice, origins of racism.

0:55

Could you give a sense of how big a topic

0:57

that is for social psychologists?

1:00

Yeah, the area that we might call

1:02

prejudice and stereotyping within social

1:04

psychology is so central

1:06

to the field that it's

1:09

obviously a very important topic to

1:11

study. But if you look at

1:13

the slice of the pie that prejudice

1:16

and stereotyping takes up compared

1:18

to the total number of topics in social

1:20

psychology, it's pretty big and probably

1:23

bigger than a lot of things

1:26

that we experience more

1:28

on a daily basis. And

1:30

it's traditionally been the case. I think that

1:33

social psychologists, many

1:35

of them view themselves as,

1:37

among

1:38

other things, trying to directly

1:41

understand social ills

1:44

and in the hopes of being able

1:46

to change society. This is

1:49

true from the earliest of social psychologists

1:51

who came from Europe and were

1:53

interested in understanding what had happened

1:56

in World War II, what had happened in Germany that

1:58

had enabled people to turn. on

2:00

Jews and on other groups in such

2:02

an ugly way. And social

2:04

psychologists also just have for many

2:07

years focused on racism,

2:10

stereotyping, prejudice, all

2:12

kinds of in-group and out-group

2:15

aggression. And

2:17

one of the things which is surprising, I

2:19

think, to people from outside

2:21

the field is that when we talk about stereotyping,

2:24

there's definitely an ugly side of it. But

2:27

part of it just is the

2:29

normal way we think about things. So

2:31

one thing I raise in my intro

2:33

psych class, and maybe the students are surprised

2:35

by this, is that categorizing

2:38

people and making inferences based on

2:40

the categories isn't some weird

2:43

pathology or act of cruelty or habit

2:45

we need to get out of. It's fundamentally

2:47

how our minds work. So

2:50

cognitive psychologists, the people who

2:52

study language and memory and rationality

2:54

and development, will often also

2:57

study how we categorize the world. And part of

2:59

categorizing the world is developing

3:02

an understanding of certain kinds of things.

3:04

So if you know what a dog is, you

3:06

probably know it typically has four legs

3:09

and a tail and barks and eats

3:12

meat and so on, you know what an apple is, you

3:14

know that you can eat it, it typically has

3:16

a certain color and so on. And

3:19

when we come to

3:20

social groups, it's not like all this

3:22

shuts down. So you

3:24

have not only just categories

3:26

of stereotypes of dogs and

3:29

apples, you have them of men and women. And

3:31

so you might notice that

3:33

men are on average taller than

3:35

women. When I was raised, I

3:37

would think, well, doctors tend to

3:39

be men and nurses tend to be women.

3:42

And we have all of these stereotypes

3:45

in our heads. You say

3:47

to somebody about a category

3:49

like Canadian or Italian

3:52

or somebody who's Jewish. And they have

3:54

a bunch of beliefs and ideas about what

3:56

members of that category should be like and call out a stereotype.

3:59

I do remember.

5:59

and being told since he's over 80

6:02

years old, he has to pay extra.

6:05

And he says, well, I'm a good driver. How could they do this

6:07

to me? My sons who are on the

6:10

young side have to pay extra also. So

6:13

there's a sort of stereotyping based on age,

6:15

which we're kind of comfortable with. We kind

6:17

of live with. But, you know, on the other hand,

6:20

I think it'd be quite a different thing if

6:22

somebody said, oh, because you're ethnicity, we're

6:25

gonna charge you more.

6:26

Even if the insurance companies that look

6:28

at our data, people of this ethnicity get

6:30

into more crashes. So we're gonna charge you. You might

6:32

just say, well, you can't do that. Right.

6:36

The social psychologist Phil Tetlock illustrates

6:39

this nicely when he talks about these

6:42

taboos. You're allowed, for

6:44

instance, to say, by

6:46

allowed here, I mean, it's deemed socially

6:49

not inappropriate to say that teenage

6:51

boys should be charged higher insurance

6:54

because statistics show that

6:56

they get into more accidents.

6:59

But you wouldn't be allowed to if the

7:01

data were that since

7:03

I'm South American, I'll just say, use my own ethnicity,

7:06

that Latin Americans get into

7:08

more accidents. Even if the numbers

7:11

said that it would be considered immoral.

7:13

Well, we'll just talk about Latin Americans. And Jews.

7:16

In this intersection. So that it's

7:18

entirely self-reflexing.

7:21

And so, right, there's an episode of the

7:23

comedy show, Curb Your Enthusiasm, where

7:26

Larry is the main character. Larry's getting divorced. And

7:28

he discovers to his shock that his lawyer

7:30

is not Jewish, as he thought he was. And

7:33

he gets another lawyer. He wants a Jewish lawyer. Jewish

7:35

lawyers are much better, he

7:37

believes, whether that's right or wrong. That's

7:39

a stereotype, which would be taboo

7:42

in some places. But

7:43

one thing is we use them all the time.

7:46

So if you're looking for directions,

7:48

you're not gonna ask a four-year-old,

7:51

because, you know, they're typically not good

7:53

at directions. If you're in New York

7:56

City and you bump into,

7:58

say, a young black gay woman.

7:59

You may assume she has

8:02

more liberal politics than most. You

8:04

could be wrong. She could be

8:06

a very right-wing person.

8:09

But odds are, given the math,

8:12

you're safe to assume her politics. And

8:14

it's quite different if you were to deal with,

8:17

you know, an older man, for instance.

8:19

These things track, they track age, they

8:22

track race, they track gender. And

8:24

it's an interesting question. What generalizations

8:27

are correct? And when is it morally

8:29

appropriate to draw upon them?

8:30

Categorizing by race does feel

8:33

a little bit different than, say, categorizing

8:36

cognitive ability by age. Race,

8:39

as a category of thing, just is

8:41

a little weird and squishy and not what

8:44

we think it is often. So that's

8:46

one of the interesting findings

8:48

in psychology. It goes under the broad term

8:51

of essentialism. And many

8:53

psychologists, Susan Gelman,

8:56

sent some wonderful developmental work on this, believe

8:59

that we essentialize categories,

9:00

and particularly social categories.

9:03

And this can have different meanings,

9:05

but the main claim here is that

9:07

we tend to think that these categories have

9:10

something intrinsic about

9:12

them that make them what they are.

9:15

And often it comes of a belief that membership

9:17

in the category is all or nothing. And

9:20

this simply isn't true.

9:22

For some Jews

9:24

and some non-Jews, to be Jewish is

9:28

to have a single parent of Jewish

9:30

ethnicity. And for

9:32

others, it has to be the mother.

9:35

And it's not like you ask a biologist

9:37

who's right. These are social

9:39

religious choices. In the United States,

9:43

there's essentially a one-drop rule

9:45

for who's black,

9:46

where if you have black

9:49

ancestry, they count you as black, even

9:52

if the majority of your genetic

9:54

ancestry is non-black.

9:57

And again, this isn't like a fact

9:59

of science.

9:59

These are social choices. It's

10:02

social choices to view, uh, something

10:05

like Barack Obama,

10:06

who's, you know, one parent is black, one parent

10:09

is white, to figure out how to categorize

10:11

and say, oh, you gotta be one or the other.

10:14

And in some countries and cultures, that

10:16

decision might go another way.

10:18

These categories aren't reflecting anything

10:21

that can be seen under a microscope. I think

10:24

that a lot of people confuse this because race

10:27

and ethnicity is obviously

10:29

based on

10:31

some sort of genetic relationship, some

10:33

sort of genetic family history, but

10:36

it's so complex the way that we use that

10:38

basic information and turn them into

10:41

categories, where we lump

10:43

an Ethiopian and a Kenyan in the

10:46

same race, even though genetically

10:48

they might be further apart than, uh,

10:50

say, a Latin American and an Asian person. That's

10:53

right. And you see tremendous social forces

10:56

at work for these decisions. You know, there's

10:58

arguments right now, are Jews white?

11:00

And whatever that argument is,

11:02

it's not an argument that's going to be resolved

11:04

by carefully studying DNA and

11:07

doing, doing careful science on it. These,

11:09

these are social,

11:10

these are social questions. And race

11:13

is, like you said, an on

11:16

combination of, to some extent, genetics.

11:19

People describe so-called races

11:21

as extended families. And there's some, there's

11:24

some truth to that, but also social

11:26

conditions as to cutoffs and

11:28

boundaries. And you

11:30

see a sort of mismatch between a

11:32

proper conceptual understanding

11:35

of these issues and how people naturally view it, where

11:38

people naturally often draw sharp

11:40

divisions where none exist. So

11:43

can we say that all of the

11:45

cognitive machinery that is

11:47

used to categorize everything

11:50

in the physical world, and

11:53

even in the world of cultural artifacts, get

11:55

recruited to make distinctions

11:58

on fairly arbitrary.

11:59

social categories. Yeah,

12:02

there's actually a huge literature

12:05

in social psychology making

12:07

that point very well. So Henri

12:09

Tajvall, who was one of the refugees

12:12

from Europe at the end of World

12:15

War II, I think, did some lovely

12:17

studies where he

12:19

wanted to study intergroup

12:22

animosity and why we hate one group over

12:24

another, but he didn't want to use real groups. He wanted

12:26

to push the point that

12:28

these groups can be as arbitrary

12:29

as you want. So he did this

12:32

wonderful study where he got people and

12:34

he gave them a sort of sham test. And in

12:36

the spirit of social psychologists everywhere, he lied

12:38

to them and told half of them that

12:41

they are Kandinsky lovers and half of them that

12:43

they are Klee lovers. Two

12:46

artists, by the way.

12:49

Two artists.

12:51

We assume everybody is intimately related

12:53

to work of Kandinsky and Klee. Friends,

12:56

we frantically Google. These were very different

12:58

times. These were

12:59

different times. That's right. That's right.

13:02

Now, that's right. Now, to be

13:04

lovers of, you know, Breaking Bad

13:06

versus Better Clueso or something.

13:09

But the finding is based on this arbitrary

13:12

distinction, people would align

13:14

themselves with the group they were put

13:16

into. The Klee lovers would

13:19

favor

13:20

other Klee lovers when they had to make a choice. They might

13:22

think Klee lovers are better. Other work

13:24

simply flips the coin. Head or tails.

13:26

I flip head, you flip tails. And we favor

13:29

our own group. There's work

13:31

with children. And I love to talk about some of

13:34

the kid work. The kid work is fascinating. But where

13:36

they put kids in red

13:38

t-shirts or blue t-shirts

13:40

and with the slightest prod,

13:42

they start organizing themselves

13:45

into different groups. And I

13:47

think this shows us something important,

13:50

which is that

13:52

we are very groupy. We are very

13:55

inclined to break the world up into us versus

13:57

them. And the pattern patterns

14:00

that we see when we chop up the world

14:02

into Jew, non-Jew,

14:05

black, white,

14:06

and all this, to some extent, reflect

14:08

just our natural propensity to find groups

14:11

everywhere. In some cases,

14:13

the groups may have some sort of genetic backing. In

14:16

other cases, like, I don't know, Capricorn versus

14:18

Sagittarius, they have none at all. There's arbitrary,

14:21

but we're quite happy to chop up the world that way.

14:23

So, let's talk about the developmental work because

14:26

one of the questions that you can ask, you

14:28

could say, well, we

14:30

know from observation and we know from

14:32

decades of social psychology work that

14:35

race

14:36

is central,

14:40

at least in some cultures, but in the United States

14:42

for sure, race

14:45

is a central

14:47

category that we use

14:49

to make inferences

14:52

about other people. And

14:54

some of it just depends on visual cues, how

14:57

I categorize somebody as belonging

14:59

to one race or another, based on how they look, whether

15:01

that's correct or incorrect. But it's

15:04

so central, that way of

15:07

categorizing people into races and ethnicity

15:09

seems so central, that it might

15:11

seem like it's a natural way.

15:14

Like,

15:15

some people might even think, well, we've

15:17

evolved to look at skin color

15:19

or facial features as a way to distinguish

15:21

out-group versus in-group members, which

15:24

I think has some problems. But

15:27

one way to answer the question is to look at kids.

15:29

And so, do kids, do very

15:31

young children categorize people based

15:34

on this appearance? It's

15:37

a great question. I did some work

15:39

on babies' appreciation

15:42

of in-groups and out-groups with my colleagues

15:44

at Yale.

15:45

And this ended up getting some popular

15:47

press attention and I see it, you

15:49

know, online as babies born

15:51

bigots or babies are little racists.

15:54

And I don't think that that's right. I think it's more complicated.

15:57

So from the get-go, babies...

15:59

babies are subject

16:01

to in-group, out-group biases. That is, if

16:04

you put them in different categories, what

16:06

t-shirt they're wearing or something like that, they

16:08

will favor their own category. Very

16:11

early on, they break up the world by age.

16:13

They distinguish children from adults. Very

16:16

early on, they break up the world by gender,

16:18

distinguishing boys from girls,

16:21

men from women. Very early

16:23

on, and this is a real discovery from

16:26

Katie Kinsler and her colleagues, they're

16:29

very focused on language.

16:31

So there's some nice experiments

16:33

where you have an American kid say a

16:35

baby or a two-year-old or a three-year-old say

16:38

a black woman

16:40

is handing the baby, offering the baby

16:42

some food. A white woman is

16:45

also offering the baby some food. And

16:48

one of them is speaking with an accent.

16:50

Babies ignore color altogether, and

16:52

they avoid the one speaking with an accent.

16:55

English-speaking babies prefer English

16:57

speakers with no accent. French-speaking

17:00

babies who are exposed to French prefer

17:02

French speakers without an accent. Now,

17:04

race is really complicated. So here's,

17:07

there's two things to keep in mind. One

17:09

thing is babies do show a preference

17:12

for looking at faces that are the same

17:14

race as the faces they

17:17

have been exposed to. There's a familiarity

17:19

bias.

17:20

So it's not the kid's own skin

17:22

color. The kid probably doesn't even know his or

17:24

her own skin color. But a

17:27

kid raised by light-skinned people would

17:29

prefer to look at light-skinned

17:30

people. Raised by dark-skinned people

17:33

prefer to look at dark-skinned people. A

17:35

study done with Ethiopian immigrants

17:37

to Israel, where the kids were raised by

17:39

multi-colored people, found

17:42

no preference. Kids preferred to look at everybody.

17:45

So people saw this initial data and said, oh,

17:47

they're kind of biased. But it turns

17:49

out that when it comes to their choices, who

17:52

to interact with, who to take a toy

17:54

from, who to want to talk to. Bias

17:57

based on what we call race

17:59

up until quite later.

18:01

It's like kids notice the difference and

18:03

it affects who they like to look at, but they

18:05

don't take it seriously. It's only

18:08

later when the society

18:11

says, you know, you know, the color of your skin, that's

18:13

important.

18:14

That's how we divvy up resources. That's

18:16

who we decide who plays with whom. Then kids

18:18

develop biases. So kids raised,

18:21

for instance, in schools where

18:23

there just isn't much

18:25

racial distinction. Race

18:28

isn't made a big deal of 10 not to develop

18:30

any racist biases. Kids raised in

18:32

schools where race is a big thing do develop

18:35

racist biases. It's like you start

18:37

off with a natural tendency

18:39

to break the world up into us versus them.

18:41

And then society tells you who's

18:43

to us and who's to

18:44

them. I alluded to something

18:46

earlier when I was talking about whether or not

18:48

there were

18:50

evolutionary accounts that might

18:54

predict that infants

18:56

would

18:57

care about race. And I said they're problematic

18:59

because race is not something

19:02

that our ancestors

19:04

had a notion of in the sense

19:07

that we do because they didn't come into contact, you

19:09

know, black people and Asian people

19:11

didn't meet. This doesn't

19:14

seem like it would be a category

19:16

that's built in in any way. You

19:18

would need our ancestors to have

19:20

actually encountered people who looked drastically

19:23

different from them. And that just didn't happen. But what

19:25

probably did happen was

19:27

the

19:27

local tribe spoke a different language

19:30

and that might be a very efficient

19:32

way of knowing the difference between us and them. That's

19:35

right. And it's a nice example of how evolutionary

19:37

considerations can shape

19:40

how we do our psychology. So it's

19:42

long been known that when

19:44

you look at somebody, there's three

19:47

things you take away from it, three main categories.

19:50

And their age,

19:52

gender or sex, however you want to view it, and

19:54

race. And we find this, for instance,

19:56

in memory confusion studies. So

19:59

if you...

19:59

If you hear a sentence said by an

20:02

old white woman,

20:04

you're more likely to misremember

20:07

it when you get it wrong and then attribute

20:09

it to another old white woman, as

20:12

opposed to a younger black man, for

20:14

instance. And in fact,

20:17

when people make some often embarrassing mistakes

20:19

in public, whereas they mix up one

20:22

person with another, what

20:24

makes it embarrassing is they typically get the race

20:26

right and the age

20:29

right and the gender right, and then

20:31

just the person wrong. But

20:33

the evolutionary psychologists

20:34

say, OK, age makes

20:37

sense. From a standpoint

20:39

of dealing with a person, it really matters whether

20:41

you're dealing with a five-year-old, a 30-year-old,

20:43

or a 60-year-old. It

20:45

determines how much. Are

20:49

you going to get a fight with them? Do you have to care

20:51

for them? Are they going to be healthy?

20:54

Sex slash gender matters a lot.

20:57

That's who you're going to mate with, how

21:01

it relates to family structures. But race,

21:03

for exactly the reason you're saying, is

21:06

a puzzle. Why in the world do we take race so

21:08

seriously when it's not something we've

21:10

been attuned to in our evolutionary past?

21:13

And the solution is what we alluded to before,

21:15

which is that we don't take

21:17

race seriously in and of itself.

21:20

We take race seriously insofar

21:23

as it's a surrogate

21:24

for social

21:27

categories that matter.

21:29

So if somebody was to say,

21:31

in this sense, race is a social construction, I

21:33

think there's a lot of truth to it, which is societies

21:36

for different reasons make genetic

21:39

ancestry and historical categories

21:42

count a lot. And you race in that

21:44

society and say, wow, where I'm living,

21:46

the in group is this race and the

21:49

out group is that race.

21:50

Yeah. And you point to

21:52

a study that was done a number of years

21:55

ago on race really

21:57

is a proxy for coalitions.

21:59

along the lines of what you're saying, where

22:03

they used people of different races, but

22:05

wearing shirts representing

22:08

different teams.

22:09

That's right. So the analogy

22:12

here is you're watching a

22:14

basketball game. And some

22:17

players on one team are

22:20

white, some are black, some players on the other team are

22:22

white, others are black. But there

22:24

you're watching, you're seeing

22:27

a real powerful case of two groups in competition.

22:30

In that case, when you think of it that way,

22:33

race disappears, you no longer think of

22:35

them in terms of race. And so there

22:38

in a memory confusion problem,

22:41

you're then

22:42

more likely to confuse say

22:45

a black player from team

22:47

A with a white player from team

22:49

A than with a black player from team

22:52

B.

22:53

All of a sudden, you don't care about race

22:55

because you're now in a situation where race

22:58

doesn't matter.

22:59

And a lot of the sort

23:02

of practical side of social psychology

23:04

says, one way to

23:07

fix strong racial biases is

23:10

to bring people into

23:12

situations where race just

23:14

matters less. It's sometimes

23:16

called the contact hypothesis where

23:19

if you meet people from different races and different

23:21

ethnicities, in a situation

23:24

which is egalitarian and you're all fighting

23:26

for a common cause and everything, like sports,

23:28

like the military,

23:29

race will just matter less.

23:32

What that illustrates is that this

23:34

kind of categorizing and

23:37

being prejudiced on the basis of race can be

23:40

erased as the authors call it. But

23:43

it also shows how important

23:46

race as a category in our culture

23:48

is. We're constantly in

23:51

a condition where it's not sports, it's

23:54

race and ethnicity.

23:55

That's right. And there's long been a political

23:58

debate over, how do you do? deal

24:00

with it. Where

24:01

one moral

24:03

stance says we should aspire

24:05

towards a sort of colorblindness, where

24:08

race doesn't matter at all.

24:10

And that's a goal, both personal goal and

24:12

a societal goal. Other people have a different

24:14

perspective.

24:15

And they say, no, we should actually just focus

24:18

more on our racial identities,

24:21

make sure the different groups treat one another

24:23

with respect. But

24:26

we shouldn't in some way try to erase race.

24:28

And in fact, you

24:30

might, for instance, argue that trying to erase

24:33

race in the here and

24:36

now will just serve

24:38

to perpetrate pre-existing inequalities

24:41

without addressing them. But

24:43

these are questions that are, I wouldn't

24:45

say above our pay grade, but they're,

24:48

again, they're political ones, they're not psychological

24:50

ones. Yeah, that's right. The very question

24:52

of what to do given that

24:55

members of certain races have

24:57

been treated poorly for hundreds of

24:59

years, it seems like a very facile

25:01

solution to say, oh, let's stop talking about

25:03

it. Let's

25:06

just forget about race for now. And

25:09

there's the fact that as psychologists,

25:11

we tend to talk about race and ethnicity, we

25:13

use it interchangeably, in

25:15

terms of the evils of racism and

25:17

discrimination and cruelty. But

25:20

in the real world, people identify

25:23

themselves with the different groups. They

25:25

take pride in them. They

25:27

take them seriously.

25:28

I think many

25:31

people, if you just told them, you

25:33

know, we have a plan and place to obliterate

25:35

your group, not by killing you all,

25:38

but simply so that in the future, don't we

25:40

know more people who call themselves Jews?

25:42

Nobody, no more people who call themselves Italians

25:45

would say, well, I'd like, I like my group.

25:48

I think my group adds some stuff. I identify

25:50

myself with them. And that's a consideration

25:53

to be taken into account as

25:55

well. This episode

25:57

is brought to you once again by Give Well.

26:00

If you're considering donating to charity,

26:02

or if you've never donated to charity before and you're

26:05

interested in figuring out where your

26:07

money would go, well, that can

26:09

be difficult, if not downright impossible.

26:11

Charities aren't exactly

26:14

open and forthcoming without information

26:16

sometimes on their websites. So it

26:18

can be hard to track down exactly what

26:21

your money is doing and how much impact it will

26:23

have. That's what GiveWell is all

26:25

about. They've dedicated a ton

26:28

of staff, over 40,000 hours each year,

26:32

to trying to find the highest

26:34

impact, evidence-backed opportunities

26:37

for people to donate their hard-earned

26:40

money. Over 100,000 donors

26:42

have used GiveWell's research

26:44

and their service to donate more

26:46

than $1 billion. And

26:48

rigorous evidence suggests that these donations will save

26:51

over 150,000 lives and

26:53

improve the lives of millions more.

26:56

So if you are just

26:58

interested in the research, you can go to GiveWell.org and

27:01

they've published all of their research absolutely

27:04

free. You don't even need to sign up. No personal

27:06

information needs to be given. And you can

27:08

use all of their research to figure out where

27:10

to donate to, where

27:13

your money will go, and essentially

27:15

how much impact each dollar that you give will

27:17

have. But also, if

27:19

you wanna use GiveWell and

27:21

donate through them, they'll allocate

27:24

your tax-deductible donation to the charity or

27:26

fund of your choice, and they won't take a cut

27:28

at all. They're not about taking

27:31

any money. They really just want

27:33

to make the process of

27:35

donating to charity more effective. So

27:38

if you've never donated to GiveWell's recommended

27:40

charities before, you can have your donation

27:42

matched up to $500 before

27:45

the end of the year, or as long

27:47

as matching funds last. To claim

27:49

your match, all you have to do is go to GiveWell.org

27:53

and pick podcast and enter psych

27:55

at checkout, just to make sure that

27:57

they know you heard about them through us.

27:59

GiveWell.org, pick podcast

28:02

and enter Psych at checkout. Our thanks

28:04

to GiveWell for sponsoring this episode

28:06

of Psych. So

28:09

a lot of what we're talking about here,

28:12

and this is how sort of people tend to talk about

28:14

about race and is really explicit biases.

28:16

You know, you ask me, who do you want to date?

28:19

What do you think of the blue t-shirt

28:22

people? What do you think of the red t-shirt people? You

28:24

know, tell me honestly your

28:26

stereotypes about about the Jews. But

28:30

psychologists, social psychologists, and

28:32

this is one of the the big revolutions

28:34

in the field, have long been interested

28:36

in implicit

28:38

biases, biases that.

28:41

Well, I'll let you define it, because the

28:43

definition, the definition is controversial. What

28:46

is what is an implicit bias? Let

28:48

me start with a little bit of history of social

28:50

psychology. Used to be the case when

28:52

social psychology started that if you wanted to know

28:55

whether or not somebody was prejudiced, if

28:57

you want to know whether a white American had

28:59

negative attitudes toward black Americans,

29:01

you just ask them

29:03

and they would tell you. If

29:06

you said, would you let a black person

29:09

date your daughter? They would just say no

29:11

if they were prejudiced. Obviously,

29:14

after a while, that stopped being

29:16

the case. Now, as many social

29:19

psychologists and

29:21

most people suspected, this wasn't

29:23

because prejudice went away entirely.

29:26

It was simply because people

29:28

knew that it was a bad thing to express

29:31

such preferences.

29:32

Now, the question became,

29:34

has it gotten to a point where

29:39

it's just that people are lying about

29:41

these racial prejudices?

29:44

Maybe it's the case

29:46

that

29:47

even when people are being honest and

29:49

saying, I have no such preferences,

29:52

there is a deeper part of them

29:54

that still harbors those prejudices.

29:57

These are not things that are even expressed.

29:59

Explicitly available to their own

30:02

awareness. So if you ask the average person

30:04

are you racist they would say no and

30:06

they wouldn't be lying

30:08

But

30:10

you look at society and you say

30:13

Clearly there still is a Large

30:17

amount of inequality based on

30:19

on race and ethnicity So

30:21

maybe what's going on is that

30:24

people don't themselves know that

30:26

they are biased in this way Nonetheless

30:30

in the depths

30:32

of their mind they're still harboring

30:35

these prejudices and this might come

30:37

out say in their behavior or as

30:41

social psychologists Discovered

30:44

they could do in some tasks

30:47

that are Meant to assess

30:50

your attitudes in a way that you

30:52

have no real control over So

30:55

so just to jump in there are sort of three categories

30:58

three ways to be racist though

30:59

whether the word racist should

31:01

apply for all room is unclear one way is

31:04

that I I Don't

31:06

want my daughter to date Italians

31:09

and I will tell you that

31:10

the second one is that I

31:13

don't want my daughter to date Italians and I

31:15

I'll keep it secret But

31:18

I know I have this belief the third

31:20

the most interesting sort is I had

31:23

this animosity against Italians But I don't know

31:25

it myself

31:27

At some level I feel I say

31:29

I'm not prejudiced at all and I'm

31:30

honest about it.

31:32

I'm not I'm not hiding anything but

31:35

in subtle ways my

31:37

bias shows through

31:39

and we call that implicit bias, that's

31:42

right and and I think

31:45

That

31:45

there is a lay sense that this happens Like

31:49

let's let's go back to Freud Freud

31:51

was interested in what he called parapraxis slips

31:53

of the tongue You might imagine that somebody

31:55

says no problem with Italians.

31:58

It's fine

31:59

But every once in a while, a little comment

32:02

emerges that illustrates that

32:05

maybe their view of Italians is

32:07

negative. They're more quickly to

32:09

think of Italians when they think of a

32:12

particular trait that's nasty.

32:16

An off-color joke about Italians

32:19

escapes them, even though they insist

32:21

that they're not at all biased.

32:23

It's Italians, they just think it's a funny joke. It

32:26

is very Freudian. It is very Freudian.

32:28

In fact, some of the

32:30

examples could come from Freud.

32:33

For instance, one way a bias will show itself

32:35

would be,

32:37

suppose I want

32:40

to describe a typical social psychologist. And

32:43

I say, well, he.

32:44

And I describe a man.

32:47

And what if most people do that? That

32:49

would suggest, but if you ask me,

32:52

do I think a social psychologist has to be

32:54

a man? Of course not. Maybe if you

32:56

ask me and even say, oh, there's just as many women

32:58

in the field as men. Certainly for developmental

33:00

psychologists, there's more women

33:02

than men. But maybe what comes to mind

33:05

is a man. And

33:08

that's the sort of implicit system working.

33:13

You know, the extreme example is, you know, I'm thinking

33:15

about a criminal.

33:17

And I say, oh, I could... And automatically,

33:20

a certain race comes to mind. I'm thinking of a

33:23

brave hero and a certain race comes

33:25

to mind.

33:26

And so social psychologists who are

33:28

interested in the

33:31

possibility that

33:34

prejudice

33:35

exists, it's lurking

33:38

deeper in the mind than

33:40

we might think, develop some

33:43

tools based on some work

33:45

in cognitive psychology, develop tools

33:47

to try to assess whether or not this was the case.

33:51

And one of them that you highlight

33:52

was developed by our friend and

33:54

former colleague, Mazarim Banaji and Tony

33:56

Greenwald. It's called the Implicit Associations

33:58

Test.

33:59

Do you want to describe that? I don't know if we should.

34:02

Yeah. Yeah. Yeah. OK. So

34:04

here's one sort of example.

34:07

Suppose you take a test looking at whether

34:09

or not you're biased towards the elderly

34:12

or against the elderly. And you ask me, oh, I have

34:14

no bias at all. I don't care about age. But

34:17

then you play a little game where they're

34:19

on a computer screen. And you say you're going to see

34:22

faces of old people

34:24

and young people.

34:25

And if you see, and you're

34:27

also going to see words, and the word's going

34:29

to be positive, like good

34:32

or wonderful, or negative,

34:34

like poison.

34:35

And here's what you have to do. In

34:38

one condition, people are told, when

34:40

you see an old face or

34:42

a positive word, hit this button.

34:45

And if you see a young face and a negative

34:47

word, hit that button so they get a series

34:49

of

34:50

faces and words and a pounding

34:52

away button. And how fast they do it

34:56

is measured. In the other condition,

34:59

they're told, if you see an old face and a negative

35:01

word, hit this button. Versus

35:03

if you see a young face and a positive word,

35:06

hit that button.

35:07

And in this subtle way,

35:10

you kind of tap whether or

35:12

not there is an association between

35:14

old and positive or young

35:16

and positive. And in fact, what this study

35:18

will find is people are a little bit slower

35:22

matching old and positive

35:24

and young and negative than

35:26

they are matching old and negative

35:29

and young and positive,

35:30

presumably because positivity

35:34

is linked more up to the young and

35:36

negativity is linked more up to the old.

35:39

That a fair description? That was great. And

35:42

you can very nicely give

35:44

people these tests and ask them

35:47

before or after, explicitly, do

35:49

you hold any bias against old people? And

35:52

you can compare their explicit

35:54

and their implicit attitudes. And

35:57

what you find with many of these categories

35:59

about race.

35:59

about age is that there is

36:02

a dissociation. People say,

36:04

no,

36:05

I don't have any bias. Nonetheless,

36:07

this implicit task demonstrates

36:10

a bias. So the real question

36:12

is what's going on here? Is

36:14

this showing that in fact people

36:17

are being prejudiced or they do in

36:20

fact hold prejudice attitudes?

36:23

What does it even mean if they don't know that

36:25

they do? Yeah. And here you

36:28

and I may have, may have different takes. I'm somewhat

36:30

skeptical about what these studies show. I'll

36:33

first say what I'm not skeptical

36:35

about, which is I think in the aggregate,

36:37

a lot of these studies can show some

36:40

very interesting things, often

36:42

about relative bias. So I

36:45

talk about a study by Mazarim

36:47

Banaji with Tessa Charlesworth.

36:50

And what they did was they took online data from 4.5

36:52

million subjects from 2007

36:55

to 2016. And

36:58

they looked at their attitudes about sexual orientation,

37:00

race, skin tone, age, disability

37:03

and body weight.

37:04

And they also asked about explicit

37:06

attitudes like, you know, I strongly

37:08

prefer young people to old people. What do you

37:10

think of that? And what they found

37:12

was over these, about

37:14

this about 10 year period, people

37:17

became less biased explicitly for

37:19

just about everything.

37:21

And, you know, that's maybe not a huge surprise.

37:23

You know, you could tell, say, say regarding

37:25

gay people, there's been profound differences

37:28

in attitudes towards gay people over those

37:30

times. But the implicit

37:32

attitudes also found a decline

37:35

in some ways. There's a decline in

37:37

implicit negative attitudes for race, skin tone,

37:40

sexual orientation, less of a decline

37:42

for age and disability, and our attitudes

37:44

got worse in the domain of

37:46

body weight, suggesting that at least implicitly,

37:49

we become harder

37:51

on people who deviate

37:53

from the ideal weight. And

37:55

so it's really interesting to use these measures

37:57

to say, you know, well, you know, we see,

37:59

uh,

37:59

this sort of attitude gets worse, this sort of attitude

38:02

gets better. That's what I like

38:04

about these studies. I mean, there's other studies

38:06

using sort of simpler methods that

38:09

compare how seriously, if

38:11

you're evaluating somebody's resume and

38:13

there's a black face on it

38:14

versus a white face, how biased

38:17

are you? What if it's somebody who describes

38:19

those conservative or

38:21

liberal? How biased are you?

38:23

I think that stuff's really cool. Here's my

38:25

criticism. Here's my problem, set of problems.

38:27

These are often used

38:30

as an individual

38:32

measure of how racist you are.

38:34

I take the test, you could take the test too,

38:37

you go online, you just Google Harvard

38:39

IAT and it'll pop up.

38:41

And this test will show

38:43

me in which ways I have implicit biases

38:46

towards one group or another. And

38:49

then sometimes people walk away saying, oh

38:51

my God, I'm racist. I'm sexist.

38:54

I don't like gay people.

38:56

I have all these biases. And

38:59

I think this is the wrong moral

39:01

to take from it for a couple of reasons. And

39:04

to be fair, the people who do these

39:06

tests are now very cautious against

39:08

using them as individual measures of how

39:10

racist you are. So one point

39:13

is there's a lot of variability

39:15

in the responses. If

39:17

I don't like how I perform on the test

39:20

now, I could take it again tomorrow and

39:22

my results will change quite a bit. A

39:25

second thing is that there's not

39:27

much of a correlation

39:29

between how well you score on these

39:32

tests and how

39:34

you behave in the real world. There's some controversy

39:37

over this, but when big studies

39:39

are done, it turns out that doing

39:41

very badly on a race IAT doesn't

39:44

seem to predict that you're particularly nasty

39:47

towards people of a different race. And

39:49

then the third concern, which is more theoretical,

39:51

is it's not clear this measures animus,

39:54

like really dislike or hatred,

39:56

which you might think is part and parcel really

39:58

being racist.

39:59

as opposed to just associations.

40:02

So, you know, if

40:04

I score, if I find it quicker to

40:08

relate elderly people with bad words

40:10

and then I do to write them to good words,

40:13

that might not mean that I myself have anything against

40:15

elderly people. It might just mean

40:18

that I understand in the world, the world we

40:20

live in,

40:21

that there are biases against elderly

40:23

people.

40:24

And so too for people who weigh

40:26

more than average and so on. For

40:29

all these reasons, I'm cautious about

40:31

going too far in what the

40:34

IT shows us. I am cautious as well.

40:37

And I do disagree and I'll lay out

40:39

a bit about why I disagree. But

40:42

I think what we do agree on is

40:44

that certainly the case

40:46

hasn't yet been made in

40:49

any convincing way that

40:51

these measures as individual

40:53

differences

40:54

ought to be treated as showing whether

40:57

or not you're truly racist. And

40:59

I think one of the errors that

41:01

the researchers made early on was

41:04

in giving individual feedback. So

41:06

I remember taking the IT for the first

41:09

time and getting

41:11

the score,

41:13

telling me that I am kind of medium anti-black.

41:17

That can throw people for a loop. In fact,

41:20

the feedback that many African-Americans

41:22

received was that they were anti-black.

41:26

And this was upsetting

41:28

for many. I

41:31

don't think we

41:32

ought to use these tests for the

41:34

problems that you outlined as an assessment

41:37

of deep or true attitudes.

41:40

Especially not if you go the route of believing

41:42

that what's going on is that you're

41:45

just lying about your explicit

41:47

attitudes. And now we've discovered

41:50

that you in fact are racist. And

41:52

one of the things that people have

41:54

tried,

41:55

flirted with let's say,

41:58

is using these kinds of... tests

42:02

to make hiring decisions. Yeah. Which

42:05

I think is sorely

42:07

misguided. If the

42:09

only criticism were the first

42:12

one that you mentioned that what we

42:14

call test retest reliability

42:16

is low,

42:17

is the chances that your score tomorrow

42:20

will be the same as your score today is

42:23

bad. It would be like having a scale

42:25

that gave you a different weight from

42:27

day to day without anything

42:29

having changed in your actual

42:32

weight. We would just call that a terrible measure.

42:34

We would say that's a bad scale. Let's

42:36

toss it out. We need to get a better scale. And

42:39

just to jump in just on one subtle

42:41

point, which is you might think that we're sort of contradicting

42:44

one another, where I say these tests in

42:46

the aggregate

42:47

could tell us something about changes through

42:50

time. And now we're saying, but

42:52

they're so unreliable. But these

42:55

two could both be right.

42:56

So your scale analogy is very good. Suppose

42:59

your scale could be off by five pounds, either.

43:01

Either way. So that's that's a horrible

43:03

way of measuring how much I weigh.

43:06

But if you use that lousy

43:09

scale and you found that from

43:11

the years, you know, from the last last

43:13

decade, people's weight went gradually

43:16

up.

43:17

That could still be a real finding. Even

43:20

an unreliable measure, if it's unreliable

43:22

within certain parameters, can tell you about broader

43:24

patterns. That's right. I mean, for that. Yeah,

43:27

no, that's a it's a great point. And

43:30

and think of it as taking a driving

43:32

exam when you have to pass the test for

43:34

your driver's license.

43:36

One day you might be nervous. You might

43:38

not have slept well. Your performance

43:40

will go down for whatever reason. The

43:42

instructor wasn't

43:45

very kind to you, made you nervous.

43:48

Next time you take it, your driving ability

43:51

hasn't changed, but you got better sleep.

43:53

You had less caffeine.

43:55

The driving instructor was kind and polite

43:58

and you perform better. It could be that this

44:01

has nothing to do with your true

44:03

driving ability, but rather just some

44:05

feature of the environment that was causing

44:08

your score to change. If

44:10

you took enough driving tests, if

44:12

you went for 14 days

44:15

in a row and we computed the average score, we

44:19

would have something meaningful

44:22

because

44:23

over all of those trials, you

44:25

would start being able to statistically

44:28

see a pattern. Whereas, just taking

44:30

it once is bad. So just taking

44:33

the implicit association test once I think

44:35

is not going to tell you that much, even

44:38

though we've just told you to go to the implicit website.

44:42

You have to know that this is the case. You

44:45

might take it again tomorrow and it would be different. But

44:48

in what we just said in whether you aggregate

44:51

thousands of people over decades, um,

44:54

you could aggregate one person

44:57

over multiple, multiple trials, just

44:59

like the driving test example. And

45:02

if this test is measuring anything

45:05

of, of importance, what you might get

45:07

is, uh, a more

45:09

reliable estimate of

45:11

what your quote unquote true implicit attitudes

45:14

are. So

45:16

we could get rid of the error that,

45:19

that one day, um, something

45:21

happened that made you score differently than another

45:23

day. So let's say that that's the case.

45:25

Let's take care of that statistical issue.

45:27

The, the second thing that you brought up was

45:29

that research has

45:31

not done a good job of showing

45:34

that your score on this thing

45:36

is good at predicting real world prejudice

45:40

behavior. And you mentioned one

45:42

reason might be that it's just a bad measure, right?

45:45

Which might account for all of these findings

45:47

that it's, if it's unreliable, if you only took

45:49

it once, it has no real shot

45:52

of predicting what your real world

45:54

prejudice behavior would be. But suppose we had

45:56

a better measure. We

45:58

gave it to you a bunch of times and we. a better

46:00

estimate. Would that predict your real

46:03

world precious behavior? And I think here you

46:05

have to look at the way in which we

46:07

try to assess real world precious behavior.

46:09

Maybe a true bigot, like

46:12

a real just nasty

46:14

white supremacist person who

46:17

goes around saying bad

46:20

things about black people

46:22

and who is doing things

46:24

like refusing to hire

46:27

them based on their race or telling their children

46:29

to never date them. That

46:32

would give you a lot of real world behaviors.

46:36

And that would give you again,

46:38

a reliable estimate

46:41

of whether or not this person was in fact

46:43

behaving like this in the real world. But

46:45

it's difficult in the range

46:47

of people that we're talking about, your

46:50

average person, how can

46:53

you tell from their behaviors,

46:55

whether or not they're precious? Suppose that

46:58

one time in the entire year, they

47:01

refuse to hire someone on the basis of their race.

47:04

I would say, well, that's a bigoted thing to do.

47:06

And I would even call that person

47:08

racist.

47:09

That ways in which we try

47:11

to assess real world behavior are

47:14

bad.

47:15

They only give us the smallest shot

47:18

of a thin slice of an individual's life,

47:21

given the situation we put them in and given

47:23

the measures that we're using. We don't

47:26

have a very good chance of documenting

47:29

a real world, precious behavior. And

47:32

I think for that reason, we

47:34

need to think not

47:36

that these studies are showing that there

47:39

is no relationship, rather they're just

47:41

failing to show anything because you got

47:43

bad measures on both sides. So

47:46

if you followed someone

47:47

and

47:48

documented all of their behaviors over

47:50

a month, you had lots

47:53

and lots of observations of their behavior.

47:56

Then maybe I think you could come up with

47:58

a reliable estimate of whether

47:59

or not their behavior showed prejudice in any

48:02

way. And then if you match that with a very

48:04

good estimate of their implicit attitudes, there

48:07

you finally might get a shot of

48:09

answering the question, do your implicit attitudes

48:11

predict your prejudice behavior? And we haven't

48:14

done that. No,

48:14

that's fair enough.

48:16

But let me ask. So

48:18

you do this, you test an individual

48:20

a hundred times, get a good

48:22

sense of them. And then you monitor,

48:25

maybe there's a camera on them for

48:27

a full year and a team of people

48:30

judges how well they treat people of different

48:32

races. And you say, and suppose you're right,

48:35

you do get a relationship between your implicit

48:37

attitude and how racist

48:39

they are in the real world. And just to be clear, I'm not saying

48:41

that you would, I'm saying that we have zero

48:43

evidence really, I think. Right. So

48:46

it's the absence of evidence isn't

48:48

evidence of absence. I'm saying that we haven't

48:50

found much predictive power. You say it hasn't been put to the

48:52

test properly. But do you

48:55

think it would do better than asking

48:57

this, this, you know, yeah, I'm now imagining

49:00

my own stereotypes are and I'm imagining a

49:02

southern, you know, sweaty sheriff,

49:05

you know, chomping a cigar white guy, you know,

49:08

like the dukes of hazard. I'm imagining

49:10

that the sheriff and the dukes of hazard as your prototypical

49:13

racist. And there, there my stereotypes are

49:15

at work.

49:16

Do you think you would do

49:18

better off with this implicit measure as opposed

49:21

to asking him an anonymous poll?

49:24

Tell me, what do you think of black people?

49:26

No, I think in those cases,

49:28

the implicit measure wouldn't be adding

49:31

much. I think the only interesting cases

49:33

are where you have the dissociation

49:36

and what you're, you're presumably trying

49:38

to uncover is that, is that

49:42

there is something that people

49:44

either deny or don't even know. Because

49:48

I think in those, in those cases, you

49:50

know, if you, if you show me pictures of snakes

49:52

and spiders versus flowers, the

49:55

implicit association test does a very good job

49:57

showing that I associate negatives

49:58

with snakes. snakes and spiders and

50:01

positives with flowers. And

50:03

so

50:04

you could have just asked me. You're getting

50:06

no additional. In fact, for racism,

50:09

if they're willing to say it out

50:11

loud, then that is probably

50:14

a more reliable metric. So

50:16

you're more an reasonist than me. Let me raise

50:19

some. You tell me if this is a legitimate worry. I

50:21

can see two types of people doing

50:24

very poorly, that is showing a strong

50:26

racist response on the race IAT.

50:29

One is this racist guy who,

50:31

whenever he thinks of a black person thinks, oh,

50:34

as awful, I think of a white person

50:36

that's the best. But the other

50:38

sort of person I could see doing poorly is

50:41

maybe somebody who's himself is a

50:44

black person who's extremely steeped

50:46

in knowledge and appreciation of

50:48

discrimination

50:50

against black people. So they

50:52

see black linked up with

50:55

a black face linked up with horrible on the

50:57

IAT. Oh God, that goes together so

50:59

well, because there's so many people think they're horrible.

51:01

And so there they're responding,

51:04

not based on their own animus,

51:07

but rather on their appreciation of

51:09

prejudice exists in the environment. Somebody

51:12

who is, for instance, to change the example

51:14

of it, who is very upset

51:16

about discrimination against people who are overweight,

51:19

may be really quick to associate

51:22

overweight people with negative words, because

51:24

that's exactly what they're upset about. So they see the connection.

51:28

Yeah, that's a good point. And I hadn't addressed your

51:30

third concern that maybe, which

51:33

is what this example is showing, that

51:36

maybe there are people who simply

51:38

have

51:39

these associations in

51:41

their minds because they exist in the

51:43

world so much. And so the person who

51:45

is really in depth studying the

51:48

problem of anti-fat prejudice

51:51

will have read so many things that may be

51:53

fat and bad or concepts that are linked up

51:56

so strongly. That's a

51:58

good question. It's an empirical question.

51:59

that

52:01

I don't know if we've answered sufficiently.

52:04

But let me push back a little bit on this, because

52:07

it kind of gets

52:09

to what I think is the heart of this debate,

52:11

which is, what do we mean

52:14

by having a believer having an attitude? Yes. So

52:18

in many domains, you would say,

52:21

what does it mean to learn something? We

52:24

talked about this a bit with behaviorism. You

52:27

learn something by associative

52:30

processes. This thing and this thing

52:32

go together. In our episode

52:34

on behaviorism, I told you the story about

52:36

how, via classical

52:39

conditioning, I came to associate the HBO

52:41

logo, the little screen

52:44

that comes up with the static and the sound of

52:46

static and the HBO visual.

52:50

Because it was attached to the Sopranos,

52:53

it came right before the Sopranos, and I came

52:55

to love the Sopranos, I developed

52:57

this connection between the

53:00

static HBO logo and

53:03

good, solely through it being

53:06

consistently paired with a good

53:08

thing. In that case, we would just say

53:10

that I have a good attitude toward

53:13

the static.

53:15

Sure, I never endorsed

53:17

it. I never even really thought

53:19

about it. But just the fact

53:22

that over and over again, it's been

53:24

associated with something that I deem to be good,

53:27

makes me think that what

53:29

I have is a pro attitude toward

53:32

that logo. Why is it the case

53:34

that if you've shown me a black

53:36

man and violence, a black man and violence,

53:38

a black man and violence over and over again,

53:41

and now I have this

53:43

association that

53:45

is,

53:46

when I see a black male face,

53:49

I think negative and violence, how

53:51

is that not an attitude? I

53:54

think that this

53:55

critique of implicit association

53:58

tests as tracking associative, associations

54:00

and therefore not really being belief

54:02

is Mistaking the

54:04

external world for the internal one. It's

54:07

only interesting

54:09

that The associations are

54:11

out in the world because they've

54:14

made their way into my mind

54:17

Once they've made their way into my mind. How

54:19

can I distinguish my black

54:21

male and violence? Connection

54:24

from my HBO logo and

54:26

good vibes. Yeah, that's a good point

54:29

I'll say I'll say two things in response One

54:31

thing is the HBO logo example

54:33

cuts both ways because not only

54:35

do you associate it with a feeling of happiness?

54:38

But you know you do. Yeah,

54:40

so there is you got the implicit

54:42

but at the same time the explicit comes for the ride

54:44

But Paul I think that there might have been a moment when

54:47

I didn't know

54:48

Maybe and there was a moment of realization

54:50

when that logo came up in a completely unrelated

54:53

show that I was feeling excited And

54:55

then I reasoned my way to Oh

54:58

That's right. I have this association with this

55:00

logo No That's fair enough And

55:02

it's also it's also fair enough

55:05

that sometimes you put together these things

55:07

in your head and it seems weird to say it's

55:09

just an association Well if it carries

55:11

negative feelings with it or positive

55:13

feelings Then you know we

55:16

could argue about whether the word attitude, but

55:18

it has emotional weight to it, right? It's

55:20

not this abstract thing All I'll say

55:22

is there's is research suggesting that emotional

55:24

weight isn't necessary to get it in

55:27

your head So these experiments

55:28

where they don't give you

55:30

the actual expert experience. They just tell

55:32

you they say let me tell you about two made-up

55:35

groups You know at crime in

55:37

a faraway land You know the flobbles

55:40

are despised for their bad behavior

55:42

and the bubbles are beloved and everybody loves

55:44

the bubbles And then you give them an IAT with

55:46

flobbles and wobbles and boom

55:49

It gets treated like a despised

55:51

group and a love group in the real world even though

55:53

you just just heard this It's just stuff

55:56

you stored in your head Yeah,

55:58

I mean I think we're converging on ideas

55:59

that there could be

56:01

a heterogeneity of many different

56:03

types of implicit

56:06

cognitions, some emotionally late

56:08

and some not.

56:09

I would fully endorse

56:11

that. What I would say about those cases is

56:15

there is something negative there, but

56:18

it's probably just not

56:20

that powerful. But once you get, I guess

56:23

what I believe is once you get the evaluative

56:25

system kicking in, once you

56:27

have paired something, categorize

56:31

it as negative, that there is

56:33

some weight there. Now, whether that's psychologically

56:36

relevant, and in

56:38

the sense that this would make a difference

56:41

to my expressed attitudes or my behaviors,

56:44

might be way, way too weak. I mean, I

56:46

don't think that there's any realistic

56:48

way in which you could say that

56:51

I'm

56:52

racist against flobbles or whatever

56:55

the group was. But I will say this,

56:57

and I'm going to be very candid here, because

57:00

I think it's important to express

57:03

my own experience with this stuff. Having

57:05

been raised as an, you

57:08

know, my family were immigrants, came

57:10

to the United States as Latin Americans, I

57:13

identify as Hispanic, but I have

57:15

been raised in this

57:17

country with all of the racist

57:20

ills that we know

57:23

that it has. I lived in Miami,

57:25

then California. Two different

57:27

times I saw race riots happen

57:30

in cities very close to me. I'm

57:32

not immune from the negative

57:35

associations of especially

57:37

black men and bad. And

57:40

when I took the IAT for the first

57:42

time, which wasn't on the website, it was in

57:44

Mazarim Banaji's lab, as they

57:46

had recently developed it. And

57:48

I got the result that I had

57:50

some negative associations

57:53

with black faces. I wasn't surprised

57:55

at all.

57:56

I'm a little racist and

57:58

I have zero

57:59

a problem admitting that. And

58:02

it's not just against black people. I'm a little

58:04

racist about a lot of groups. It's

58:06

something that I know because I get

58:10

those little pangs of negative

58:13

thoughts,

58:14

negative feelings in certain

58:16

situations, not in all of them. And

58:19

I think that it's deeply important for my

58:21

own development as a moral

58:23

human being that I understand

58:26

that those things happen because the only

58:29

chance I'll have of correcting them or

58:31

at least trying to correct them is

58:33

by first realizing that I have them. And

58:36

in that sense, I think the IAT, as

58:39

crappy as it is, is a psychometric

58:41

test, might be giving us

58:43

something valuable, which is you could imagine

58:46

that somebody might live their life in denial by

58:48

saying, there is nothing in me. There's

58:50

not a prejudice bone in my body. When

58:53

in reality, they're having mild

58:56

negative attitudes towards a whole lot of individuals

58:59

and groups that might

59:02

leak out

59:02

and might be noticed by members of those groups,

59:05

but not at all even by the person who's

59:08

leaking them out.

59:09

So in that way, you're saying

59:11

one of the powers of IAT is it could

59:13

make what would otherwise be unconscious

59:16

and bring it up to your consciousness, bring up the conscious

59:19

awareness. Like Freud said, the goal of therapy

59:21

to make the unconscious conscious. It

59:24

gives you insight. And

59:27

it's not really magic. I see

59:29

your point, which is that every

59:31

time the HBO logo

59:33

comes on it, you give a little smile. And

59:36

I say to you, boy, you've really associated that with

59:39

sopranos.

59:39

That makes you right. You're right.

59:41

Now,

59:41

when

59:44

I reflect on it, I see that. And

59:46

the dark side is I say,

59:49

we've tested you and you seem to

59:51

associate this group with negativity.

59:53

And you say, yeah,

59:55

I see that. And then

59:57

the project, which is the project.

59:59

I end a chapter on this project is, what

1:00:02

do you do about it? I mean, suppose

1:00:04

when I'm facing a member

1:00:06

of a certain group, I'm interviewing them,

1:00:09

I am biased against them. I'm biased

1:00:11

against them because they're black. I'm biased

1:00:13

against them because they're a man. And

1:00:16

most of my people I interview for developmental

1:00:18

psychology graduate are women. I'm

1:00:21

biased against them because they don't share my political

1:00:23

views, maybe because they're

1:00:25

trans or gay, maybe because they're

1:00:28

overweight, maybe because they're physically unattractive.

1:00:29

Two things which are powerful

1:00:32

biases against, we have a

1:00:34

lot of look at them. So one step

1:00:36

is I kind of know it. And knowing it, I

1:00:38

might say, look, I'm trying my

1:00:40

best to give them a fair, fairly

1:00:43

interview them and so on. But I have to be aware

1:00:45

of the possibility I'm gonna be additionally hard on

1:00:47

them. I'm gonna say, oh,

1:00:50

they're too aggressive, they're

1:00:52

not very personable, or their answers aren't as

1:00:54

smart. Well, in the magical

1:00:56

world, we could switch all those buttons.

1:00:59

I wouldn't say the same thing.

1:00:59

And so I leave the interview

1:01:02

saying, based on purely

1:01:05

objective facts,

1:01:07

I'm not gonna hire the person, but it weren't objective

1:01:10

at all. They just felt objective. They were percolated

1:01:12

through my biases, implicit or explicit.

1:01:15

So

1:01:16

how do we fix that? And

1:01:19

I'll try out an argument, which is, knowledge

1:01:23

of the problem is part of the thing, but

1:01:25

simply trying hard is

1:01:27

not gonna work.

1:01:28

You can't make bias go away

1:01:31

by just saying, I'm not going to be biased.

1:01:34

Take a more mild, acceptable example

1:01:36

of bias. If I'm reading essays

1:01:39

and judging how good they are, and one of them

1:01:41

is by one of my sons, I'm

1:01:43

gonna say, well, I'm not gonna be biased. I'm

1:01:46

gonna read his fairly. And then I'm gonna

1:01:48

come up to you and say, this is the best essay. I wasn't

1:01:50

biased at all. I tried not to be biased, I succeeded. It's

1:01:52

just a damn best essay. And

1:01:55

you wouldn't trust me.

1:01:56

You would say, you failed.

1:02:00

So I think the solution for these

1:02:02

things, and I'm not the first to make this argument, is

1:02:05

you don't change the person's head. You

1:02:08

change the world in certain ways. So if

1:02:10

I'm biased by how attractive a person is, I

1:02:13

should make my judgments without seeing how they

1:02:15

look. If I'm biased by their skin color, I

1:02:17

should not know their skin color.

1:02:19

If I'm biased by my relative

1:02:21

or not, I shouldn't know that.

1:02:24

So one solution is to do all these things blind

1:02:27

to the circumstances that

1:02:29

might bias you. Another very different

1:02:31

thing is to have some sort of thing like quotas

1:02:33

or requirements that kind of force

1:02:36

an equity to counterbalance

1:02:40

our own biases that we don't like.

1:02:42

And they're very different solutions

1:02:44

to do as a colorblind versus they have

1:02:46

a rule experts then have to be of this group. And

1:02:49

they come from very different political commitments, but

1:02:51

they both serve the same purpose,

1:02:54

which is how do you get rid of a bias

1:02:56

that people can't shake just by wanting to?

1:02:59

Yeah, I think it's absolutely right

1:03:01

that we can't rely on just our sheer

1:03:03

will to undo these biases. There's,

1:03:06

as far as I know, there is no bias in all

1:03:08

of psychology for which that works.

1:03:12

It is a notoriously

1:03:14

difficult and thorny issue, and

1:03:16

particularly in this case, because

1:03:19

as you say, the, and

1:03:20

as Freud might take delight in us

1:03:22

saying, we can rationalize our way

1:03:26

through a whole lot of prejudice behaviors.

1:03:28

So I agree with everything you said. And in fact, if

1:03:30

we want to see large scale changes, making

1:03:33

changes to the structures, to the rules

1:03:36

is, I think the best and perhaps

1:03:38

only way to go. I just want to add

1:03:41

something though. And that thing that I want to

1:03:43

add is more about our individual

1:03:46

moral development. There is

1:03:48

virtue in making

1:03:51

systemic structural changes for

1:03:53

the sake of equity. That probably

1:03:55

makes a way bigger difference. But

1:03:58

if it comes

1:03:58

to my own.

1:03:59

views and my own prejudices.

1:04:02

Sure, I can't will my way into

1:04:05

being less prejudiced, but

1:04:07

what I can do is

1:04:10

expose myself and build relationships

1:04:13

with people who are members of

1:04:15

the groups that I might have a slightly

1:04:18

negative or implicit bias against. And

1:04:20

I think in my own life that's been the one

1:04:24

thing that has been a salve

1:04:28

for my own racism, for my own

1:04:30

prejudices, is to develop relationships

1:04:33

with people and

1:04:35

actually love those people and actually

1:04:37

care for those people. And

1:04:40

there I feel

1:04:42

my prejudices kind of

1:04:45

simmering away, kind of falling

1:04:47

away in a bit. And I think

1:04:49

that I've become a better person. So

1:04:52

you don't like Mexicans, go become

1:04:54

friends, go to a community

1:04:57

event, make friends

1:05:00

with people who are Mexican. And

1:05:02

I think that not only will

1:05:04

you change, maybe you're changed

1:05:07

on the IT score, you will become a better

1:05:09

person because you

1:05:11

will have exposed yourself to

1:05:13

and made yourself even care

1:05:16

about and love people who you

1:05:19

might never have done otherwise. So

1:05:22

this is my Pollyannaish view of

1:05:24

how to defeat prejudice. It

1:05:27

might not work at a societal level, but

1:05:29

I look at it like this. If you want to

1:05:31

stop poverty, the best thing probably do

1:05:34

is to donate to one of the many

1:05:36

charities. Work hard. If you're an attorney,

1:05:38

you work hard, you make $500 an hour, work

1:05:41

harder and make more money and give

1:05:43

it to that charity. Sure. But

1:05:45

if you want to change yourself, volunteering

1:05:48

to help out people in your community

1:05:51

who are poor

1:05:52

and actually spending some time with

1:05:54

them,

1:05:55

I think

1:05:56

will make a change

1:05:59

in your life and probably in the lives of those people

1:06:01

that would never have come about with the more cold,

1:06:04

rational, systemic change. How's

1:06:07

that? Wow. You've

1:06:10

come out as a much kinder

1:06:12

person than I am. I was

1:06:15

just satisfied to make systemic social

1:06:17

changes and then not work on

1:06:19

myself so much, but I'm finding this

1:06:22

inspiring. Well,

1:06:25

all I'll say is besides you being an excellent role

1:06:27

model for me, because I think those are

1:06:29

excellent points. And there's

1:06:31

broader debates over effective altruism

1:06:34

and the general question of how to be a good person, where

1:06:37

I think you're reminding us

1:06:39

and reminding me that often

1:06:41

sort of thinking, how can I make the world a better place?

1:06:45

We should also be thinking, how do we

1:06:47

make ourselves better people within this world?

1:06:50

What I'll say is that there are many people who do exactly

1:06:52

the opposite

1:06:54

of what you're describing. You

1:06:56

can find anything on social media and

1:06:59

there's a lot of people I know who

1:07:01

delight, and that's the word I

1:07:03

would use delight in showing videos

1:07:05

of say black people committing crimes. And

1:07:08

they watch them over and over again. And they're

1:07:10

doing the opposite. They're purposefully

1:07:13

hardening their hearts. And

1:07:15

we do that when we are in a

1:07:19

sort of confrontation with a group. Your

1:07:21

country's at war. You'll watch atrocities

1:07:24

being done by the other side. You're in a

1:07:26

political battle. You'll watch

1:07:29

the other side, be they Democrats, Republicans,

1:07:31

or whatever. You'll watch your ugliest moments

1:07:34

and you'll try to build up your stereotypes,

1:07:37

your biases, implicit and explicit

1:07:40

in ways that

1:07:42

expand on your hatred.

1:07:44

And I think that's a very human trait. And I think,

1:07:47

I don't know what you're suggesting is

1:07:49

a very moral alternative.

1:08:00

E E

1:08:04

E E

1:08:08

E E

1:08:12

E E

1:08:16

E E Bulsar

1:08:22

E E

1:08:26

E E