Press "Enter" to skip to content

How the **** does memory work? | Matthew Kelley | TEDxLakeForestCollege


I really struggled putting together a

I really struggled putting together a title for today’s talk so I had to asked

title for today’s talk so I had to asked

title for today’s talk so I had to asked my wife for help

my wife for help

my wife for help fortunately for me she’s an

fortunately for me she’s an

fortunately for me she’s an award-winning novelist and it took her

award-winning novelist and it took her

award-winning novelist and it took her all of three seconds to say you need a

all of three seconds to say you need a

all of three seconds to say you need a title

title

title I got your title this is your site I

I got your title this is your site I

I got your title this is your site I have the same passion I loved it but I

have the same passion I loved it but I

have the same passion I loved it but I worried me a little bit after all this

worried me a little bit after all this

worried me a little bit after all this is a TED talk so I thought I’d soften it

is a TED talk so I thought I’d soften it

is a TED talk so I thought I’d soften it a bit maybe go with this one but I was

a bit maybe go with this one but I was

a bit maybe go with this one but I was still concerned so in the end I settled

still concerned so in the end I settled

still concerned so in the end I settled on this last one that’s an interesting

on this last one that’s an interesting

on this last one that’s an interesting thing about these titles is that none of

thing about these titles is that none of

thing about these titles is that none of the words presented are overtly vulgar

the words presented are overtly vulgar

the words presented are overtly vulgar or profane but if you loved me you

or profane but if you loved me you

or profane but if you loved me you filled in this missing information with

filled in this missing information with

filled in this missing information with a word that could be considered taboo

a word that could be considered taboo

a word that could be considered taboo now

now

now cognitive psychologists many of them

cognitive psychologists many of them

cognitive psychologists many of them study this generation process how do we

study this generation process how do we

study this generation process how do we go about filling in this missing

go about filling in this missing

go about filling in this missing information but I’m actually interested

information but I’m actually interested

information but I’m actually interested in the next step once we’ve generated

in the next step once we’ve generated

in the next step once we’ve generated something how does that influence our

something how does that influence our

something how does that influence our later memory researchers have explored

later memory researchers have explored

later memory researchers have explored the effect of generation on memory in

the effect of generation on memory in

the effect of generation on memory in the lab since the late 1970s typically

the lab since the late 1970s typically

the lab since the late 1970s typically using a procedure like this participant

using a procedure like this participant

using a procedure like this participant comes into the lab and they’re asked to

comes into the lab and they’re asked to

comes into the lab and they’re asked to read a long list of words maybe 20 or 30

read a long list of words maybe 20 or 30

read a long list of words maybe 20 or 30 words though resemble this some of the

words though resemble this some of the

words though resemble this some of the words are presented in their full intact

words are presented in their full intact

words are presented in their full intact form so participants simply need to read

form so participants simply need to read

form so participants simply need to read them aloud as they see them others are

them aloud as they see them others are

them aloud as they see them others are presented as word fragments with one

presented as word fragments with one

presented as word fragments with one letter missing before participants can

letter missing before participants can

letter missing before participants can read these aloud they have to fill in

read these aloud they have to fill in

read these aloud they have to fill in the missing letter but this tends to be

the missing letter but this tends to be

the missing letter but this tends to be trivially easy and thoughts we could

trivially easy and thoughts we could

trivially easy and thoughts we could demonstrate it right now but I need your

demonstrate it right now but I need your

demonstrate it right now but I need your help

help

help what is the second word

what is the second word

what is the second word village fourth word Wow word fantastic

village fourth word Wow word fantastic

village fourth word Wow word fantastic despite the relative ease of reading and

despite the relative ease of reading and

despite the relative ease of reading and generating these words participants tend

generating these words participants tend

generating these words participants tend to remember significantly more of the

to remember significantly more of the

to remember significantly more of the fragmented or generated words compared

fragmented or generated words compared

fragmented or generated words compared to the intact or read words the simple

to the intact or read words the simple

to the intact or read words the simple act of filling in a missing letter can

act of filling in a missing letter can

act of filling in a missing letter can boost your memory by 10 to 30 percent

boost your memory by 10 to 30 percent

boost your memory by 10 to 30 percent depending on the conditions in the study

depending on the conditions in the study

depending on the conditions in the study so why does this happen well it seems

so why does this happen well it seems

so why does this happen well it seems that generation encourages more semantic

that generation encourages more semantic

that generation encourages more semantic or meaning based processing during

or meaning based processing during

or meaning based processing during encoding which then we additionally for

encoding which then we additionally for

encoding which then we additionally for you to retrieve later on at tests now if

you to retrieve later on at tests now if

you to retrieve later on at tests now if you’re like most of my students you’re

you’re like most of my students you’re

you’re like most of my students you’re probably thinking this is interesting

probably thinking this is interesting

probably thinking this is interesting but it all seems rather artificial

but it all seems rather artificial

but it all seems rather artificial alright after all talk titles typically

alright after all talk titles typically

alright after all talk titles typically don’t contain missing words and you

don’t contain missing words and you

don’t contain missing words and you rarely run into word fragments in your

rarely run into word fragments in your

rarely run into word fragments in your daily lives I guess unless you’re a fan

daily lives I guess unless you’re a fan

daily lives I guess unless you’re a fan of crossword puzzles so for the rest of

of crossword puzzles so for the rest of

of crossword puzzles so for the rest of today I want to share with you some

today I want to share with you some

today I want to share with you some research from my lab and from my

research from my lab and from my

research from my lab and from my classroom that explores two real-world

classroom that explores two real-world

classroom that explores two real-world applications of the generation effect

applications of the generation effect

applications of the generation effect and we’re going to start with an

and we’re going to start with an

and we’re going to start with an interesting project on lyrical

interesting project on lyrical

interesting project on lyrical censorship and memory you’ve most likely

censorship and memory you’ve most likely

censorship and memory you’ve most likely experienced lyrical censorship to the

experienced lyrical censorship to the

experienced lyrical censorship to the radio we’ve been there listening and

radio we’ve been there listening and

radio we’ve been there listening and then you notice that some of the lyrics

then you notice that some of the lyrics

then you notice that some of the lyrics some of the artists words have been

some of the artists words have been

some of the artists words have been omitted or obscured in some way

omitted or obscured in some way

omitted or obscured in some way regardless of the method the logic and

regardless of the method the logic and

regardless of the method the logic and the intent is clear if you can’t hear

the intent is clear if you can’t hear

the intent is clear if you can’t hear the word then you can’t process the word

the word then you can’t process the word

the word then you can’t process the word censorship is an attempt to limit your

censorship is an attempt to limit your

censorship is an attempt to limit your processing just a couple of minutes ago

processing just a couple of minutes ago

processing just a couple of minutes ago I showed you how easy it is for us to

I showed you how easy it is for us to

I showed you how easy it is for us to generate missing information wouldn’t it

generate missing information wouldn’t it

generate missing information wouldn’t it be ironic then if this simple act of

be ironic then if this simple act of

be ironic then if this simple act of censoring an item actually made that

censoring an item actually made that

censoring an item actually made that item more memorable

item more memorable

item more memorable when I made this connection I set out to

when I made this connection I set out to

when I made this connection I set out to design a series of experiments to

design a series of experiments to

design a series of experiments to explore whether we would see a

explore whether we would see a

explore whether we would see a generation effect in the context of

generation effect in the context of

generation effect in the context of lyrical censorship now all we have time

lyrical censorship now all we have time

lyrical censorship now all we have time to describe a portion of one experiment

to describe a portion of one experiment

to describe a portion of one experiment today but please know that the data are

today but please know that the data are

today but please know that the data are representative of the entire program of

representative of the entire program of

representative of the entire program of research so the first big decision I had

research so the first big decision I had

research so the first big decision I had to make in this research project was

to make in this research project was

to make in this research project was what kind of information should I censor

what kind of information should I censor

what kind of information should I censor initially I thought I’d used

initially I thought I’d used

initially I thought I’d used traditionally censored materials curse

traditionally censored materials curse

traditionally censored materials curse words drug references things of that

words drug references things of that

words drug references things of that nature but I was concerned that this

nature but I was concerned that this

nature but I was concerned that this information came from too small of a set

information came from too small of a set

information came from too small of a set it’s too well-known it might be too easy

it’s too well-known it might be too easy

it’s too well-known it might be too easy for the participants to generate and in

for the participants to generate and in

for the participants to generate and in fact this might be a class of words that

fact this might be a class of words that

fact this might be a class of words that are inherently more memorable so instead

are inherently more memorable so instead

are inherently more memorable so instead I decided to censor common nouns from a

I decided to censor common nouns from a

I decided to censor common nouns from a song I thought this would be a more

song I thought this would be a more

song I thought this would be a more difficult task for participants much

difficult task for participants much

difficult task for participants much more challenging also I figured this is

more challenging also I figured this is

more challenging also I figured this is a much larger and unknown set of

a much larger and unknown set of

a much larger and unknown set of material plus if I use nouns we have

material plus if I use nouns we have

material plus if I use nouns we have these things called normative data so I

these things called normative data so I

these things called normative data so I could ensure that all of the words were

could ensure that all of the words were

could ensure that all of the words were equally memorable at the outset of the

equally memorable at the outset of the

equally memorable at the outset of the experiment so I wouldn’t have to worry

experiment so I wouldn’t have to worry

experiment so I wouldn’t have to worry about that so with the materials in mind

about that so with the materials in mind

about that so with the materials in mind I chose a set of 30 common nouns words

I chose a set of 30 common nouns words

I chose a set of 30 common nouns words like father and chair and I wrote a song

like father and chair and I wrote a song

like father and chair and I wrote a song around them now you’re probably thinking

around them now you’re probably thinking

around them now you’re probably thinking why did he go through the trouble of

why did he go through the trouble of

why did he go through the trouble of writing his own saw and songs exists in

writing his own saw and songs exists in

writing his own saw and songs exists in the world well you see I needed to

the world well you see I needed to

the world well you see I needed to equate song familiarity across my

equate song familiarity across my

equate song familiarity across my participants which were college students

participants which were college students

participants which were college students so I couldn’t use anything too familiar

so I couldn’t use anything too familiar

so I couldn’t use anything too familiar I suppose I could have chosen an obscure

I suppose I could have chosen an obscure

I suppose I could have chosen an obscure song and obscure and but if you know

song and obscure and but if you know

song and obscure and but if you know anything about college students

anything about college students

anything about college students listening to obscure bands is kind of

listening to obscure bands is kind of

listening to obscure bands is kind of what they do so I thought that would be

what they do so I thought that would be

what they do so I thought that would be better off just designing my own

better off just designing my own

better off just designing my own experiment designing my own song

experiment designing my own song

experiment designing my own song so with the song in hand I decided to

so with the song in hand I decided to

so with the song in hand I decided to manipulate sensor

manipulate sensor

manipulate sensor in two different ways some of the

in two different ways some of the

in two different ways some of the participants heard partially censored

participants heard partially censored

participants heard partially censored information in which the initial speech

information in which the initial speech

information in which the initial speech sound was still audible you can kind of

sound was still audible you can kind of

sound was still audible you can kind of think of this as a sloppy version of

think of this as a sloppy version of

think of this as a sloppy version of censorship which actually occurs quite

censorship which actually occurs quite

censorship which actually occurs quite on the radio the other participants

on the radio the other participants

on the radio the other participants heard completely censored lyrics so all

heard completely censored lyrics so all

heard completely censored lyrics so all of the speech sounds were gone well I

of the speech sounds were gone well I

of the speech sounds were gone well I wanted to give you a sense of what this

wanted to give you a sense of what this

wanted to give you a sense of what this sounded like so I’m going to play the

sounded like so I’m going to play the

sounded like so I’m going to play the first line of the song two different

first line of the song two different

first line of the song two different ways and we’re going to start with the

ways and we’re going to start with the

ways and we’re going to start with the departure the censored version so just

departure the censored version so just

departure the censored version so just listen along it should be obvious when

listen along it should be obvious when

listen along it should be obvious when the word is censored

anyone able to figure out the identity

anyone able to figure out the identity of that missing word memory fantastic

of that missing word memory fantastic

of that missing word memory fantastic well done let me take you on a trip down

well done let me take you on a trip down

well done let me take you on a trip down memory lane here’s what it sounded like

memory lane here’s what it sounded like

memory lane here’s what it sounded like in the completely censored version

so all the sounds were gone there after

so all the sounds were gone there after hearing the song I gave participants a

hearing the song I gave participants a

hearing the song I gave participants a recognition memory test it would see a

recognition memory test it would see a

recognition memory test it would see a word on the screen and they had to

word on the screen and they had to

word on the screen and they had to decide whether that word was in this

decide whether that word was in this

decide whether that word was in this song or not and all there were 60 words

song or not and all there were 60 words

song or not and all there were 60 words in this recognition test so 60 separate

in this recognition test so 60 separate

in this recognition test so 60 separate decisions 15 of the words had actually

decisions 15 of the words had actually

decisions 15 of the words had actually been heard 15 were censored from the

been heard 15 were censored from the

been heard 15 were censored from the song and then there were 30 nouns that

song and then there were 30 nouns that

song and then there were 30 nouns that didn’t appear in the song at all they

didn’t appear in the song at all they

didn’t appear in the song at all they were distractor items so what did we

were distractor items so what did we

were distractor items so what did we find

find

find well first let me orient you to the

well first let me orient you to the

well first let me orient you to the ground on the y-axis we have percentage

ground on the y-axis we have percentage

ground on the y-axis we have percentage correct going from 0 to 100 and on the

correct going from 0 to 100 and on the

correct going from 0 to 100 and on the x-axis we have the two different types

x-axis we have the two different types

x-axis we have the two different types of censorship partial and complete we’ll

of censorship partial and complete we’ll

of censorship partial and complete we’ll begin with the heard words these were

begin with the heard words these were

begin with the heard words these were words that were actually sung in the

words that were actually sung in the

words that were actually sung in the song that actually went into their ears

song that actually went into their ears

song that actually went into their ears you can see they recognized these words

you can see they recognized these words

you can see they recognized these words rather well 65% and 61% accuracy for

rather well 65% and 61% accuracy for

rather well 65% and 61% accuracy for those two conditions but now it’s moved

those two conditions but now it’s moved

those two conditions but now it’s moved to the interesting condition the

to the interesting condition the

to the interesting condition the censored words remember these are words

censored words remember these are words

censored words remember these are words that we never heard in the song words

that we never heard in the song words

that we never heard in the song words that were either partially or completely

that were either partially or completely

that were either partially or completely omitted intuitively you might think we

omitted intuitively you might think we

omitted intuitively you might think we should have zero percent accuracy how

should have zero percent accuracy how

should have zero percent accuracy how can we remember something that wasn’t

can we remember something that wasn’t

can we remember something that wasn’t even presented but based on the

even presented but based on the

even presented but based on the generation effect

generation effect

generation effect we’re actually predicting something in

we’re actually predicting something in

we’re actually predicting something in the opposite direction we actually think

the opposite direction we actually think

the opposite direction we actually think that the generated words might be

that the generated words might be

that the generated words might be remembered better than things they

remembered better than things they

remembered better than things they actually heard that’s exactly what we

actually heard that’s exactly what we

actually heard that’s exactly what we found

found

found 88% and 80% accuracy for those two

88% and 80% accuracy for those two

88% and 80% accuracy for those two conditions respectively ironically

conditions respectively ironically

conditions respectively ironically lyrical censorship enhanced memory by

lyrical censorship enhanced memory by

lyrical censorship enhanced memory by 23% and 21% for censored words these

23% and 21% for censored words these

23% and 21% for censored words these were statistically significant

were statistically significant

were statistically significant differences now although these results

differences now although these results

differences now although these results are impressive

are impressive

are impressive please remember this was with college

please remember this was with college

please remember this was with college students

students

students our college students the target audience

our college students the target audience

our college students the target audience for censorship who’s the target audience

for censorship who’s the target audience

for censorship who’s the target audience kids children right children are the

kids children right children are the

kids children right children are the target audience now no one has yet to

target audience now no one has yet to

target audience now no one has yet to run this experiment but I predict that

run this experiment but I predict that

run this experiment but I predict that censorship will be effective for kids

censorship will be effective for kids

censorship will be effective for kids especially very young children because

especially very young children because

especially very young children because they’re not particularly good at

they’re not particularly good at

they’re not particularly good at generating information yet these are

generating information yet these are

generating information yet these are skills that develop with time and

skills that develop with time and

skills that develop with time and experience plus they might not even yet

experience plus they might not even yet

experience plus they might not even yet know the identities of those censored

know the identities of those censored

know the identities of those censored words how can you generate something if

words how can you generate something if

words how can you generate something if you don’t even know the word to begin

you don’t even know the word to begin

you don’t even know the word to begin with

with

with however lyrical censorship will become

however lyrical censorship will become

however lyrical censorship will become less and less effective over time as the

less and less effective over time as the

less and less effective over time as the kids age and as they gain more

kids age and as they gain more

kids age and as they gain more experience now so far the generation

experience now so far the generation

experience now so far the generation effect probably seems like something

effect probably seems like something

effect probably seems like something that happens to us right it seems like

that happens to us right it seems like

that happens to us right it seems like it’s this automatic process that occurs

it’s this automatic process that occurs

it’s this automatic process that occurs without our conscious intent or

without our conscious intent or

without our conscious intent or conscious awareness so this begs the

conscious awareness so this begs the

conscious awareness so this begs the question can we be more deliberate about

question can we be more deliberate about

question can we be more deliberate about this can we actively use generation to

this can we actively use generation to

this can we actively use generation to enhance our memory if we think a bit

enhance our memory if we think a bit

enhance our memory if we think a bit more broadly about what it means to

more broadly about what it means to

more broadly about what it means to generate information then we can easily

generate information then we can easily

generate information then we can easily apply this to educational settings for

apply this to educational settings for

apply this to educational settings for years in my courses I have encouraged

years in my courses I have encouraged

years in my courses I have encouraged students to take advantage of the

students to take advantage of the

students to take advantage of the generation effect as they study for

generation effect as they study for

generation effect as they study for exams I’d suggest that they generate

exams I’d suggest that they generate

exams I’d suggest that they generate definitions and examples on their own

definitions and examples on their own

definitions and examples on their own instead of trying to memorize those

instead of trying to memorize those

instead of trying to memorize those provided in the book or I’d suggest that

provided in the book or I’d suggest that

provided in the book or I’d suggest that they omit certain key words in their

they omit certain key words in their

they omit certain key words in their notes so when they’re looking at over it

notes so when they’re looking at over it

notes so when they’re looking at over it again they have to fill in that jet that

again they have to fill in that jet that

again they have to fill in that jet that missing information and generate that

missing information and generate that

missing information and generate that information surprisingly students never

information surprisingly students never

information surprisingly students never listen to me and they never take this

listen to me and they never take this

listen to me and they never take this advice and I don’t understand it’s but I

advice and I don’t understand it’s but I

advice and I don’t understand it’s but I also kind of do understand it right

also kind of do understand it right

also kind of do understand it right these are successful college students

these are successful college students

these are successful college students they have spent a lifetime developing a

they have spent a lifetime developing a

they have spent a lifetime developing a set of effective study strategies

set of effective study strategies

set of effective study strategies and there were about didn’t try

and there were about didn’t try

and there were about didn’t try something new this late in their career

something new this late in their career

something new this late in their career so if you’re a professor like me and you

so if you’re a professor like me and you

so if you’re a professor like me and you want to get your students to try

want to get your students to try

want to get your students to try something new what can you do well the

something new what can you do well the

something new what can you do well the solution is actually pretty simple you

solution is actually pretty simple you

solution is actually pretty simple you just assign it as homework and this is

just assign it as homework and this is

just assign it as homework and this is what I did in a couple of recent classes

what I did in a couple of recent classes

what I did in a couple of recent classes I’ve instituted a new semester long

I’ve instituted a new semester long

I’ve instituted a new semester long project in which they had to generate

project in which they had to generate

project in which they had to generate one multiple-choice question for each of

one multiple-choice question for each of

one multiple-choice question for each of eight chapters across the term then at

eight chapters across the term then at

eight chapters across the term then at the end of the semester I was interested

the end of the semester I was interested

the end of the semester I was interested in whether this generation activity

in whether this generation activity

in whether this generation activity influenced their performance on the

influenced their performance on the

influenced their performance on the exams so for each student I went back to

exams so for each student I went back to

exams so for each student I went back to see if there were any topics that they

see if there were any topics that they

see if there were any topics that they wrote about that were left with the

wrote about that were left with the

wrote about that were left with the topics that I tested on that is if a

topics that I tested on that is if a

topics that I tested on that is if a student wrote a question on topic X and

student wrote a question on topic X and

student wrote a question on topic X and then later on there happened to be an

then later on there happened to be an

then later on there happened to be an exam question on topic X how did they

exam question on topic X how did they

exam question on topic X how did they fare on that exam question so I compared

fare on that exam question so I compared

fare on that exam question so I compared performance in that situation to over

performance in that situation to over

performance in that situation to over all baseline performance how did they

all baseline performance how did they

all baseline performance how did they Theron all of the non generated material

Theron all of the non generated material

Theron all of the non generated material as you can see here performance in the

as you can see here performance in the

as you can see here performance in the baseline or control condition was 76%

baseline or control condition was 76%

baseline or control condition was 76% not surprisingly given what we know

not surprisingly given what we know

not surprisingly given what we know about the generation affect students

about the generation affect students

about the generation affect students fared much better

fared much better

fared much better what with the generated content by

what with the generated content by

what with the generated content by creating a multiple-choice question on a

creating a multiple-choice question on a

creating a multiple-choice question on a topic they were able to improve their

topic they were able to improve their

topic they were able to improve their performance on that topic by almost two

performance on that topic by almost two

performance on that topic by almost two full letter grades presumably the act of

full letter grades presumably the act of

full letter grades presumably the act of creating these questions encouraged more

creating these questions encouraged more

creating these questions encouraged more semantic processing the students had to

semantic processing the students had to

semantic processing the students had to think more deeply about the content and

think more deeply about the content and

think more deeply about the content and that subsequently enhanced their

that subsequently enhanced their

that subsequently enhanced their memories now I don’t want to give you

memories now I don’t want to give you

memories now I don’t want to give you the impression that this only works for

the impression that this only works for

the impression that this only works for a multiple-choice although I don’t have

a multiple-choice although I don’t have

a multiple-choice although I don’t have data on this yet the logic should hold

data on this yet the logic should hold

data on this yet the logic should hold for other types of tests as well whether

for other types of tests as well whether

for other types of tests as well whether it be short answer or essay

it be short answer or essay

it be short answer or essay so long as the generation process is

so long as the generation process is

so long as the generation process is appropriate for the kind of test that

appropriate for the kind of test that

appropriate for the kind of test that you’re taking that is if you know you

you’re taking that is if you know you

you’re taking that is if you know you have a short answer question coming up

have a short answer question coming up

have a short answer question coming up you should probably spend time

you should probably spend time

you should probably spend time generating and answering short answer

generating and answering short answer

generating and answering short answer questions you want to make sure that the

questions you want to make sure that the

questions you want to make sure that the kind of thinking you’re doing during

kind of thinking you’re doing during

kind of thinking you’re doing during study is appropriate for the kind of

study is appropriate for the kind of

study is appropriate for the kind of thinking that you’re going to be asked

thinking that you’re going to be asked

thinking that you’re going to be asked to do at tests and now I hope I have

to do at tests and now I hope I have

to do at tests and now I hope I have convinced you of the simple power of the

convinced you of the simple power of the

convinced you of the simple power of the generation effect but if there are still

generation effect but if there are still

generation effect but if there are still some skeptics in the audience I’d like

some skeptics in the audience I’d like

some skeptics in the audience I’d like to end by asking you one final question

to end by asking you one final question

to end by asking you one final question what were the three talk titles from my

what were the three talk titles from my

what were the three talk titles from my opening slide you might remember that

opening slide you might remember that

opening slide you might remember that they were all questions you might not

they were all questions you might not

they were all questions you might not remember the exact construction of the

remember the exact construction of the

remember the exact construction of the questions but I’m pretty sure you’ll

questions but I’m pretty sure you’ll

questions but I’m pretty sure you’ll remember these three items and whatever

remember these three items and whatever

remember these three items and whatever you generated for them thank you for

you generated for them thank you for

you generated for them thank you for using

using

using [Applause]

”

Be First to Comment

Leave a Reply

Your email address will not be published. Required fields are marked *