1
00:00:02,600 --> 00:00:04,933
- Welcome everyone to 
Wednesday Nite @ the Lab.

2
00:00:05,033 --> 00:00:06,366
I'm Tom Zinnen.

3
00:00:06,466 --> 00:00:09,366
I work here at the UW-Madison 
Biotechnology Center.

4
00:00:09,466 --> 00:00:12,833
I also work for the Division 
of Extension Wisconsin 4-H.

5
00:00:12,933 --> 00:00:15,733
And on behalf of those folks 
and our other co-organizers,

6
00:00:15,833 --> 00:00:19,433
PBS Wisconsin, the Wisconsin 
Alumni Association,

7
00:00:19,533 --> 00:00:21,966
and the UW-Madison 
Science Alliance,

8
00:00:22,066 --> 00:00:24,633
thanks again for coming to 
Wednesday Nite @ the Lab.

9
00:00:24,733 --> 00:00:28,400
We do this every Wednesday 
night, 50 times a year.

10
00:00:28,500 --> 00:00:32,266
Tonight, it's my pleasure to 
introduce to you Haley Vlach.

11
00:00:32,366 --> 00:00:33,566
She's a professor here

12
00:00:33,666 --> 00:00:36,200
in the Department of 
Educational Psychology

13
00:00:36,300 --> 00:00:38,666
and our School of Education.

14
00:00:38,766 --> 00:00:41,066
She was born in 
Portland, Oregon,

15
00:00:41,166 --> 00:00:43,766
and went to high school at 
Mountain View High School

16
00:00:43,866 --> 00:00:45,733
in Bend, Oregon.

17
00:00:45,833 --> 00:00:47,733
And then she went to 
Carnegie Mellon University

18
00:00:47,833 --> 00:00:50,400
in Pittsburgh and studied 
business administration

19
00:00:50,500 --> 00:00:52,166
and psychology.

20
00:00:52,266 --> 00:00:53,433
Then she moved back west

21
00:00:53,533 --> 00:00:56,066
to go to the University of 
California at Los Angeles

22
00:00:57,066 --> 00:01:01,866
for her master's degree and PhD, 
both of which in psychology.

23
00:01:01,966 --> 00:01:04,600
Tonight, she's going 
to be talking with us

24
00:01:04,700 --> 00:01:07,200
about forgetting, what it is,

25
00:01:08,900 --> 00:01:10,933
and how it helps us remember.

26
00:01:12,366 --> 00:01:15,966
Pretty fascinating topic, 
a paradox.

27
00:01:16,066 --> 00:01:19,100
Please join me in 
welcoming Haley Vlach

28
00:01:19,200 --> 00:01:20,833
to Wednesday Nite @ the Lab.

29
00:01:21,833 --> 00:01:23,366
- Thanks for that introduction.

30
00:01:23,466 --> 00:01:25,833
I'm very excited to be here

31
00:01:25,933 --> 00:01:28,633
and to speak with you all 
about some of the work

32
00:01:28,733 --> 00:01:30,966
that we've been doing in my lab.

33
00:01:31,066 --> 00:01:33,966
I'm Director of the Learning 
Cognition and Development Lab.

34
00:01:34,066 --> 00:01:35,300
And in my lab,

35
00:01:35,400 --> 00:01:38,566
we're really interested in how 
people think and learn.

36
00:01:38,666 --> 00:01:40,666
We spend a lot of time 
thinking and learning

37
00:01:40,766 --> 00:01:42,533
about thinking and learning.

38
00:01:42,633 --> 00:01:43,766
And in particular,

39
00:01:43,866 --> 00:01:46,100
we're interested in 
cognitive development.

40
00:01:46,200 --> 00:01:50,266
That is, how does thinking 
change across the lifespan?

41
00:01:50,366 --> 00:01:52,633
We're particularly interested 
in thinking and learning

42
00:01:52,733 --> 00:01:55,900
during childhood, but we 
also study adults as well.

43
00:01:56,900 --> 00:01:59,500
And we spend a lot of our time 
thinking about memory

44
00:01:59,600 --> 00:02:02,766
and memory changes and how 
those changes in memory

45
00:02:02,866 --> 00:02:05,533
might contribute to our 
thinking and learning.

46
00:02:06,666 --> 00:02:08,666
And we're really 
interested in memory

47
00:02:08,766 --> 00:02:10,766
for a number of reasons.

48
00:02:10,866 --> 00:02:12,166
So the first reason

49
00:02:12,266 --> 00:02:14,733
is that it's theoretically 
very interesting.

50
00:02:14,833 --> 00:02:17,400
Like how does memory work?

51
00:02:17,500 --> 00:02:20,300
We take in a seemingly 
infinite amount of information

52
00:02:20,400 --> 00:02:22,233
from the world 
and somehow manage

53
00:02:22,333 --> 00:02:25,133
to decode that 
information, store it,

54
00:02:25,233 --> 00:02:27,366
and then retrieve it 
later when we need it.

55
00:02:27,466 --> 00:02:30,933
And so that process 
is still a mystery

56
00:02:31,033 --> 00:02:33,566
and is very motivating to study,

57
00:02:33,666 --> 00:02:36,566
just coming up with a 
theory of how memory works.

58
00:02:37,566 --> 00:02:41,466
Memory is also practically very 
important for many reasons.

59
00:02:41,566 --> 00:02:45,633
So first, it's important to 
our day-to-day functioning.

60
00:02:45,733 --> 00:02:47,900
So for instance, when we're 
going to the grocery store,

61
00:02:48,000 --> 00:02:50,733
we need to remember the things 
on our shopping list.

62
00:02:50,833 --> 00:02:54,366
And we often think in our mind, 
"Hmm, do I need more milk?"

63
00:02:54,466 --> 00:02:56,066
And you have to remember 
whether or not

64
00:02:56,166 --> 00:02:59,133
you have milk in the 
fridge currently.

65
00:02:59,233 --> 00:03:02,000
There's also many ways 
in which memory

66
00:03:02,100 --> 00:03:04,466
can be implied into training,

67
00:03:04,566 --> 00:03:07,300
educational and 
health interventions.

68
00:03:07,400 --> 00:03:10,633
Memory is at the core of 
allowing us to think and learn,

69
00:03:10,733 --> 00:03:12,633
as we will discuss more today.

70
00:03:12,733 --> 00:03:15,833
And so there are 
numerous applications.

71
00:03:15,933 --> 00:03:18,666
But on a more sort of 
spiritual or emotional level,

72
00:03:18,766 --> 00:03:22,533
memory is fascinating 
because it's what defines us.

73
00:03:22,633 --> 00:03:25,566
In other words, 
we are our memories.

74
00:03:25,666 --> 00:03:27,166
When we create 
our sense of self,

75
00:03:27,266 --> 00:03:29,466
we think back to all 
of our experiences,

76
00:03:29,566 --> 00:03:31,700
and what those experiences are,

77
00:03:31,800 --> 00:03:34,900
are the memories that 
we've stored in our brain.

78
00:03:35,000 --> 00:03:39,866
So memory truly serves 
at the foundation of the self.

79
00:03:39,966 --> 00:03:42,666
So there are many, many 
reasons to study memory.

80
00:03:44,366 --> 00:03:47,300
One thing that is 
very interesting

81
00:03:47,400 --> 00:03:50,300
to most people who are learning 
about memory for the first time

82
00:03:50,400 --> 00:03:53,866
is that remembering is not the 
most common part of memory,

83
00:03:53,966 --> 00:03:57,300
but instead the most 
ubiquitous process in memory

84
00:03:57,400 --> 00:03:59,433
is the fact that we forget.

85
00:04:00,600 --> 00:04:03,833
Today, what I'm going 
to do is I'm gonna first

86
00:04:03,933 --> 00:04:05,666
provide an overview 
on forgetting.

87
00:04:05,766 --> 00:04:09,166
What is it and what do 
we know about forgetting?

88
00:04:09,266 --> 00:04:10,566
After that brief introduction,

89
00:04:10,666 --> 00:04:13,400
we're gonna dive into 
talking about forgetting

90
00:04:13,500 --> 00:04:16,033
helping us to learn.

91
00:04:16,133 --> 00:04:18,233
And this is counterintuitive 
as we will discuss,

92
00:04:18,333 --> 00:04:20,166
because normally we 
characterize forgetting

93
00:04:20,266 --> 00:04:22,766
as a process that deters memory.

94
00:04:22,866 --> 00:04:25,666
And then finally, I'll end 
with some concluding thoughts

95
00:04:25,766 --> 00:04:28,633
about where we can go and how 
we can think about forgetting

96
00:04:28,733 --> 00:04:30,866
in our day-to-day lives.

97
00:04:30,966 --> 00:04:32,466
So first let's dive in,

98
00:04:32,566 --> 00:04:34,733
let's talk about 
what forgetting is

99
00:04:34,833 --> 00:04:37,700
and how scientists 
think about forgetting

100
00:04:37,800 --> 00:04:40,500
when we study it from a 
scientific perspective.

101
00:04:41,500 --> 00:04:44,800
Forgetting is operationally 
defined as the declining ability

102
00:04:44,900 --> 00:04:48,133
to retrieve information 
across time.

103
00:04:48,233 --> 00:04:49,933
And what you can see 
here in this image

104
00:04:50,033 --> 00:04:52,433
is on the X-axis is time.

105
00:04:52,533 --> 00:04:54,833
So number of days or a month,

106
00:04:54,933 --> 00:04:57,333
and then on the Y-axis, 
you see this equation,

107
00:04:57,433 --> 00:05:00,566
and this equation is 
called a saving score.

108
00:05:00,666 --> 00:05:03,400
And you can think about 
this as the percent

109
00:05:03,500 --> 00:05:05,833
that is recalled across time.

110
00:05:06,833 --> 00:05:10,066
And what you can see is that 
the nature of forgetting

111
00:05:10,166 --> 00:05:12,600
is such that soon 
after learning something,

112
00:05:12,700 --> 00:05:14,266
we have pretty 
good memory for it.

113
00:05:14,366 --> 00:05:18,800
But over time, that ability to 
retrieve the memory declines,

114
00:05:18,900 --> 00:05:23,466
and it declines according to 
what we call a forgetting curve.

115
00:05:23,566 --> 00:05:27,700
Okay, so the curve is 
fast-changing at first,

116
00:05:27,800 --> 00:05:29,800
but then over time slows down

117
00:05:29,900 --> 00:05:32,833
and gives sort of a 
U-shape to the curve.

118
00:05:34,633 --> 00:05:37,433
And we've been 
studying this process

119
00:05:37,533 --> 00:05:39,100
for a considerable 
amount of time,

120
00:05:39,200 --> 00:05:41,433
dating back to the late 1800s

121
00:05:41,533 --> 00:05:44,166
at the birth of 
psychological science.

122
00:05:44,266 --> 00:05:47,366
So because forgetting has been 
such a well-studied process,

123
00:05:47,466 --> 00:05:49,633
we know some things about it.

124
00:05:49,733 --> 00:05:52,300
And I'm gonna give you a brief 
overview of what we know.

125
00:05:52,400 --> 00:05:56,633
So first, we know that we 
forget across the lifespan.

126
00:05:56,733 --> 00:05:58,933
So forgetting happens 
in young infants,

127
00:05:59,033 --> 00:06:01,233
in adulthood, and 
in older adulthood.

128
00:06:01,333 --> 00:06:05,833
In other words, every stage of 
life, we observe forgetting.

129
00:06:06,833 --> 00:06:08,433
The next thing that we 
know about forgetting

130
00:06:08,533 --> 00:06:12,933
is that forgetting happens 
across tasks and timescales.

131
00:06:13,033 --> 00:06:14,933
And it's a very 
predictable pattern.

132
00:06:15,033 --> 00:06:18,233
And in fact, that pattern is 
what I just showed you earlier,

133
00:06:18,333 --> 00:06:20,000
the forgetting curve.

134
00:06:20,100 --> 00:06:22,666
Now in this figure, 
what I have here

135
00:06:22,766 --> 00:06:24,766
are two different graphs.

136
00:06:24,866 --> 00:06:26,833
On the left side of the screen,

137
00:06:26,933 --> 00:06:29,700
you'll see that the 
X-axis indicates time

138
00:06:29,800 --> 00:06:32,100
as a matter of seconds.

139
00:06:32,200 --> 00:06:33,766
However, on the right side,

140
00:06:33,866 --> 00:06:35,633
there's a different 
dimension of time,

141
00:06:35,733 --> 00:06:39,666
which is years, like 0 to 
50 years in particular.

142
00:06:39,766 --> 00:06:41,600
And then on the 
Y-axis for both graphs

143
00:06:41,700 --> 00:06:42,966
is the percent correct.

144
00:06:43,066 --> 00:06:45,433
That's the percentage 
remembered.

145
00:06:45,533 --> 00:06:46,666
And what we can see

146
00:06:46,766 --> 00:06:48,833
is regardless of whether 
or not the timescale

147
00:06:48,933 --> 00:06:51,600
is a matter of seconds as 
such in the left graph,

148
00:06:51,700 --> 00:06:52,900
or a matter of years,

149
00:06:53,000 --> 00:06:54,933
such as what we see 
on the right graph,

150
00:06:55,033 --> 00:06:59,000
we see the exact same pattern, 
that forgetting curve.

151
00:07:00,000 --> 00:07:02,000
Now, for those of you that 
are mathematical buffs,

152
00:07:02,100 --> 00:07:06,100
what you can see is that 
what is charted here

153
00:07:06,200 --> 00:07:07,333
is a power function.

154
00:07:07,433 --> 00:07:09,933
That's that equation at 
the top of the graphs.

155
00:07:10,033 --> 00:07:13,266
And indeed, forgetting 
follows a power function

156
00:07:13,366 --> 00:07:14,733
with a high degree of fit.

157
00:07:14,833 --> 00:07:16,300
So those are squared values,

158
00:07:16,400 --> 00:07:18,700
or what you're seeing 
is like the percentage

159
00:07:18,800 --> 00:07:21,800
at which the power function 
can account

160
00:07:21,900 --> 00:07:25,233
for the data that we 
observed via forgetting.

161
00:07:25,333 --> 00:07:27,400
And one thing that's 
fascinating about forgetting

162
00:07:27,500 --> 00:07:28,600
is that one of the things

163
00:07:28,700 --> 00:07:31,366
that we can mathematically 
model the best

164
00:07:31,466 --> 00:07:32,966
of human cognition.

165
00:07:33,066 --> 00:07:37,300
Indeed, capturing about 88% or 
90-some percent of the variance

166
00:07:37,400 --> 00:07:38,766
is very high.

167
00:07:38,866 --> 00:07:42,400
And so forgetting is very 
predictable across timescales,

168
00:07:42,500 --> 00:07:43,766
but also across tasks.

169
00:07:43,866 --> 00:07:47,200
We see the similar forgetting 
curves across tasks.

170
00:07:48,866 --> 00:07:50,733
The next thing we know 
about forgetting

171
00:07:50,833 --> 00:07:52,733
is that we can 
observe forgetting

172
00:07:52,833 --> 00:07:55,600
on multiple levels of analysis.

173
00:07:55,700 --> 00:07:57,633
In other words, 
we see forgetting

174
00:07:57,733 --> 00:08:01,300
on the level of an individual 
neuron in our brain,

175
00:08:01,400 --> 00:08:03,666
but we can also 
observe forgetting

176
00:08:03,766 --> 00:08:05,800
in complex human behaviors,

177
00:08:05,900 --> 00:08:09,466
such as those day-to-day 
functions that we engage in,

178
00:08:09,566 --> 00:08:11,800
especially more complex things

179
00:08:11,900 --> 00:08:15,100
like learning new concepts 
in the classroom.

180
00:08:15,200 --> 00:08:18,100
So we're seeing forgetting 
at multiple levels

181
00:08:18,200 --> 00:08:19,966
of our being as humans.

182
00:08:21,133 --> 00:08:23,333
We also know that 
we're not alone.

183
00:08:24,466 --> 00:08:27,200
Humans are not the 
only ones that forget.

184
00:08:27,300 --> 00:08:30,533
We can observe forgetting 
in other species,

185
00:08:30,633 --> 00:08:34,033
even organisms that don't 
have our nervous system.

186
00:08:34,133 --> 00:08:37,100
So for example, 
in very simple organisms

187
00:08:37,200 --> 00:08:38,866
that only have a 
matter of cells,

188
00:08:38,966 --> 00:08:41,866
we observe that those 
cells will forget.

189
00:08:41,966 --> 00:08:43,400
So forgetting is not something

190
00:08:43,500 --> 00:08:46,166
that's specific to 
the human experience

191
00:08:46,266 --> 00:08:47,566
or our nervous systems;

192
00:08:47,666 --> 00:08:51,833
it's much more widespread 
across organisms.

193
00:08:53,233 --> 00:08:55,233
Okay, so at this point,

194
00:08:56,466 --> 00:08:59,533
what you were probably 
thinking is that,

195
00:08:59,633 --> 00:09:00,866
"Wow, so what you're telling me

196
00:09:00,966 --> 00:09:03,900
is that we're forgetting 
everything all the time."

197
00:09:04,000 --> 00:09:05,066
And that's right.

198
00:09:05,166 --> 00:09:08,500
We are forgetting 
everything all the time.

199
00:09:08,600 --> 00:09:11,600
It is predictable 
and it's inevitable.

200
00:09:13,200 --> 00:09:15,266
So why does this happen?

201
00:09:15,366 --> 00:09:17,466
Why is this happening 
to us all the time

202
00:09:17,566 --> 00:09:18,700
in a predictable manner?

203
00:09:18,800 --> 00:09:22,100
Well, the truth is, 
is that we don't know.

204
00:09:22,200 --> 00:09:24,066
It's a scientific mystery.

205
00:09:24,166 --> 00:09:27,066
We don't know what 
causes forgetting,

206
00:09:27,166 --> 00:09:30,000
but currently, there 
are two primary theories

207
00:09:30,100 --> 00:09:32,533
about why it is that we forget.

208
00:09:33,866 --> 00:09:38,033
The first collection of 
theories focus on interference.

209
00:09:38,133 --> 00:09:40,900
In other words, 
these theories posit

210
00:09:41,000 --> 00:09:43,600
that because we're 
continually acquiring

211
00:09:43,700 --> 00:09:45,800
more and more information,

212
00:09:45,900 --> 00:09:49,133
that that new information 
blocks us from retrieving

213
00:09:49,233 --> 00:09:52,766
the old information 
in our mind and brain.

214
00:09:52,866 --> 00:09:54,833
So if we weren't to 
learn anything new,

215
00:09:54,933 --> 00:09:56,066
we wouldn't forget,

216
00:09:56,166 --> 00:09:58,200
is essentially what 
they're arguing.

217
00:09:59,566 --> 00:10:02,933
However, there's a different 
collection of theories

218
00:10:03,033 --> 00:10:07,133
that center on time and the 
properties of the universe

219
00:10:07,233 --> 00:10:09,300
with regards to energy.

220
00:10:09,400 --> 00:10:12,066
So according to these theories, 
the reason that we forget

221
00:10:12,166 --> 00:10:15,466
is that energy passes 
across biological membranes

222
00:10:15,566 --> 00:10:18,666
according to this power function 
that you've seen earlier,

223
00:10:18,766 --> 00:10:20,833
in other words, 
the forgetting curve.

224
00:10:21,833 --> 00:10:23,766
And because of that,

225
00:10:23,866 --> 00:10:26,833
we observe forgetting at 
different units of analysis,

226
00:10:26,933 --> 00:10:29,666
like a single cell, but also,

227
00:10:29,766 --> 00:10:33,633
that scales up to complex 
human behavior as well.

228
00:10:34,900 --> 00:10:39,600
Now, one issue is that we can't 
control for time and energy.

229
00:10:39,700 --> 00:10:42,133
In other words, we can't 
put people in a vacuum

230
00:10:42,233 --> 00:10:44,266
where they're not 
experiencing time

231
00:10:44,366 --> 00:10:48,766
in order to understand whether 
it's truly interference or time.

232
00:10:48,866 --> 00:10:51,566
Moreover, we can't control time

233
00:10:51,666 --> 00:10:53,366
when we're doing 
interference studies.

234
00:10:53,466 --> 00:10:58,100
So with additional learning 
comes additional time.

235
00:10:58,200 --> 00:11:00,300
And so the reason 
that it's a mystery,

236
00:11:00,400 --> 00:11:02,133
is it interference 
or is it something

237
00:11:02,233 --> 00:11:04,600
about the properties 
of the universe?

238
00:11:04,700 --> 00:11:05,833
We just may never know

239
00:11:05,933 --> 00:11:08,000
because we can't put 
ourselves in a vacuum

240
00:11:08,100 --> 00:11:11,600
in order to control 
these two variables.

241
00:11:11,700 --> 00:11:12,866
So the debate rages on,

242
00:11:12,966 --> 00:11:15,866
but we're pretty sure that 
it's either interference,

243
00:11:15,966 --> 00:11:18,633
time, or some 
combination of the two.

244
00:11:20,133 --> 00:11:23,233
Okay, so that was the 
overview of forgetting,

245
00:11:23,333 --> 00:11:25,933
what it is and what 
we know about it.

246
00:11:26,033 --> 00:11:29,633
Now I'd like to move on and 
I'd like to tell you more

247
00:11:29,733 --> 00:11:32,266
about why forgetting is 
actually a good thing,

248
00:11:32,366 --> 00:11:34,500
because at this point 
you might be wondering,

249
00:11:34,600 --> 00:11:37,633
"Well, if it's 
inevitable, predictable,

250
00:11:37,733 --> 00:11:39,633
"and gonna prevent us 
from remembering things,

251
00:11:39,733 --> 00:11:41,133
it must be bad."

252
00:11:41,233 --> 00:11:43,000
Well, it turns out that 
that's not the case.

253
00:11:43,100 --> 00:11:45,866
Forgetting can be used 
to help us remember.

254
00:11:47,233 --> 00:11:51,333
So if forgetting is inevitable, 
how do we remember?

255
00:11:51,433 --> 00:11:55,566
This has been a central research 
question for many years.

256
00:11:55,666 --> 00:11:57,800
And what researchers 
have discovered

257
00:11:57,900 --> 00:12:00,000
are that certain 
learning environments

258
00:12:00,100 --> 00:12:04,433
or conditions of the environment 
can actually improve our memory.

259
00:12:04,533 --> 00:12:07,666
They can help us to 
retrieve information.

260
00:12:07,766 --> 00:12:09,566
One of the most 
well-studied conditions

261
00:12:09,666 --> 00:12:13,033
of the learning environment 
is called spaced learning.

262
00:12:13,133 --> 00:12:15,000
And what scientists 
have observed

263
00:12:15,100 --> 00:12:18,500
is that when we distribute 
learning events across time,

264
00:12:18,600 --> 00:12:21,433
people have better memory 
for that information

265
00:12:21,533 --> 00:12:24,800
compared to it being massed 
in immediate succession.

266
00:12:26,566 --> 00:12:28,666
So here's a real life example 
of that.

267
00:12:28,766 --> 00:12:32,233
Imagine that you are 
studying for a test.

268
00:12:32,333 --> 00:12:34,133
You have one or two options:

269
00:12:34,233 --> 00:12:37,166
you could cram the night before 
and just cram, cram, cram

270
00:12:37,266 --> 00:12:39,066
right up until the test.

271
00:12:39,166 --> 00:12:41,233
Well, that behavior would be

272
00:12:41,333 --> 00:12:43,166
studying in immediate 
succession.

273
00:12:43,266 --> 00:12:45,900
We would call that 
massed learning.

274
00:12:46,000 --> 00:12:49,200
The other option is to 
distribute your learning

275
00:12:49,300 --> 00:12:51,900
during the week and study 
a little bit every day.

276
00:12:52,000 --> 00:12:53,766
So you might study a 
little bit on Monday,

277
00:12:53,866 --> 00:12:54,933
a little bit on Tuesday,

278
00:12:55,033 --> 00:12:57,700
a little bit on Wednesday 
and so on and so forth,

279
00:12:57,800 --> 00:13:00,200
building up to the test.

280
00:13:00,300 --> 00:13:02,700
And what you may 
experience if you cram

281
00:13:02,800 --> 00:13:05,700
is that you do okay on the test, 
you do well enough,

282
00:13:05,800 --> 00:13:07,300
but a few days later 
you might say,

283
00:13:07,400 --> 00:13:09,766
"Wow, I don't remember anything

284
00:13:09,866 --> 00:13:12,933
from the test that I took 
earlier in the week."

285
00:13:13,033 --> 00:13:15,033
However, if you had 
distributed your learning

286
00:13:15,133 --> 00:13:17,000
or spaced it out across time,

287
00:13:17,100 --> 00:13:20,100
you would probably 
experience more memory

288
00:13:20,200 --> 00:13:21,433
for what you studied for.

289
00:13:21,533 --> 00:13:23,733
In other words, you'd 
actually remember

290
00:13:23,833 --> 00:13:25,100
what you were studying.

291
00:13:26,100 --> 00:13:30,433
And that's a real life example 
of how spaced learning

292
00:13:30,533 --> 00:13:34,433
promotes memory to a greater 
degree than massed learning.

293
00:13:35,433 --> 00:13:38,800
In the lab, scientists 
do this kind of work

294
00:13:38,900 --> 00:13:40,400
using word lists.

295
00:13:40,500 --> 00:13:44,066
What they do is they present 
words to participants

296
00:13:44,166 --> 00:13:45,666
one at a time.

297
00:13:45,766 --> 00:13:47,633
So what you're seeing 
here on this list

298
00:13:47,733 --> 00:13:50,233
is a series of words, 
and what you'll notice

299
00:13:50,333 --> 00:13:52,766
is some of the words 
like the word cat

300
00:13:52,866 --> 00:13:54,766
are presented in 
immediate succession,

301
00:13:54,866 --> 00:13:57,033
one right after the other.

302
00:13:57,133 --> 00:14:00,100
Whereas other words on the list, 
such as the word dog,

303
00:14:00,200 --> 00:14:01,633
is distributed across time.

304
00:14:01,733 --> 00:14:02,833
You'll see the word dog,

305
00:14:02,933 --> 00:14:04,233
and then you'll see 
some other words

306
00:14:04,333 --> 00:14:06,433
and then the word dog 
will come back.

307
00:14:06,533 --> 00:14:08,866
The consistent finding 
across these studies

308
00:14:08,966 --> 00:14:10,800
that use word lists 
is that participants

309
00:14:10,900 --> 00:14:13,900
have stronger memory for the 
words distributed across time,

310
00:14:14,000 --> 00:14:17,333
like dog, relative to the 
words that are massed in time,

311
00:14:17,433 --> 00:14:18,866
like the word cat.

312
00:14:20,400 --> 00:14:23,500
We know from over a 
thousand published studies

313
00:14:23,600 --> 00:14:25,100
that spaced learning 
promotes memory.

314
00:14:25,200 --> 00:14:27,333
It's a highly 
replicable phenomenon

315
00:14:27,433 --> 00:14:29,733
that's been studied 
across many timescales

316
00:14:29,833 --> 00:14:32,100
and many contexts.

317
00:14:32,200 --> 00:14:34,200
One of the primary reasons

318
00:14:34,300 --> 00:14:37,466
that spaced learning 
supports memory

319
00:14:37,566 --> 00:14:39,900
is actually that 
we're forgetting

320
00:14:40,000 --> 00:14:41,800
during the learning events.

321
00:14:41,900 --> 00:14:45,300
So let me briefly 
describe why this happens.

322
00:14:45,400 --> 00:14:48,000
So in between each 
spaced learning event,

323
00:14:48,100 --> 00:14:49,900
you'll forget information.

324
00:14:50,000 --> 00:14:51,333
And what that forgetting does

325
00:14:51,433 --> 00:14:54,733
is it makes it harder to 
retrieve what you've learned.

326
00:14:54,833 --> 00:14:56,400
But at subsequent 
learning events,

327
00:14:56,500 --> 00:14:58,033
what you'll do is 
you'll retrieve

328
00:14:58,133 --> 00:15:01,033
your prior information 
from memory

329
00:15:01,133 --> 00:15:03,633
and the cognitive effort 
that you engage in

330
00:15:03,733 --> 00:15:06,000
in remembering that information

331
00:15:06,100 --> 00:15:09,400
will then be translated 
into a slower forgetting

332
00:15:09,500 --> 00:15:10,766
for that information.

333
00:15:10,866 --> 00:15:12,766
In other words, 
by trying to remember

334
00:15:12,866 --> 00:15:14,833
and practicing that retrieval,

335
00:15:14,933 --> 00:15:18,500
you'll then have a slower 
forgetting rate in the future.

336
00:15:18,600 --> 00:15:21,433
And so by engaging in that 
retrieval practice

337
00:15:21,533 --> 00:15:23,633
across distributed 
learning events,

338
00:15:23,733 --> 00:15:25,366
you're actually 
slowing forgetting

339
00:15:25,466 --> 00:15:27,600
and promoting your memory.

340
00:15:28,600 --> 00:15:31,100
Now, researchers have 
for a long time now

341
00:15:31,200 --> 00:15:33,100
known that spaced learning 
promotes memory

342
00:15:33,200 --> 00:15:37,000
and that forgetting can be 
manipulated to promote memory,

343
00:15:37,100 --> 00:15:41,800
but this isn't true 
for all forms of memory.

344
00:15:41,900 --> 00:15:44,100
For example, researchers 
have questioned

345
00:15:44,200 --> 00:15:46,266
whether or not this 
process of forgetting

346
00:15:46,366 --> 00:15:48,000
is good for everything.

347
00:15:49,100 --> 00:15:52,033
And in particular, 
researchers have asked,

348
00:15:52,133 --> 00:15:55,700
can forgetting promote 
categorization

349
00:15:55,800 --> 00:15:57,400
and concept learning?

350
00:15:58,733 --> 00:16:00,433
And the categorization 
and concept learning

351
00:16:00,533 --> 00:16:02,733
is learning that things 
share similarities

352
00:16:02,833 --> 00:16:03,900
and differences in the world.

353
00:16:04,000 --> 00:16:07,633
So for example, a golden 
retriever is called a dog

354
00:16:07,733 --> 00:16:09,900
and a wiener dog is 
also called a dog.

355
00:16:10,000 --> 00:16:11,233
Even though they 
look very different,

356
00:16:11,333 --> 00:16:13,533
they share similarities 
and thus belong

357
00:16:13,633 --> 00:16:15,533
to the category of dog.

358
00:16:16,966 --> 00:16:18,700
And general intuition would say,

359
00:16:18,800 --> 00:16:21,100
"Hmm, I don't know if 
forgetting is good,"

360
00:16:21,200 --> 00:16:23,000
because in general, 
intuition tells us

361
00:16:23,100 --> 00:16:25,933
that forgetting is not good, 
that it's a bad thing.

362
00:16:26,033 --> 00:16:29,033
However, even researchers 
who are experts in this field

363
00:16:29,133 --> 00:16:31,133
really argued that yes,

364
00:16:31,233 --> 00:16:33,633
forgetting should 
deter categorization

365
00:16:33,733 --> 00:16:35,500
in concept learning.

366
00:16:35,600 --> 00:16:39,500
It's really fun to go back 
through the previous literature

367
00:16:39,600 --> 00:16:41,600
and read old studies

368
00:16:41,700 --> 00:16:43,500
because there's some 
fighting words out there.

369
00:16:43,600 --> 00:16:45,333
People really had strong beliefs

370
00:16:45,433 --> 00:16:48,866
that forgetting would deter 
conceptual development.

371
00:16:48,966 --> 00:16:51,100
My favorite quote is 
from Ernie Rothkopf,

372
00:16:51,200 --> 00:16:52,733
where he argued that 
things like forgetting

373
00:16:52,833 --> 00:16:55,400
and spaced learning would 
be the "friend of recall,"

374
00:16:55,500 --> 00:16:59,000
so help us to remember, 
"but the enemy of induction."

375
00:16:59,100 --> 00:17:01,933
So in other words, it would 
deter our ability

376
00:17:02,033 --> 00:17:04,033
to learn categories 
or learn concepts

377
00:17:04,133 --> 00:17:07,700
where we have to abstract out 
across our experiences.

378
00:17:07,800 --> 00:17:10,466
Let me walk you 
through this argument.

379
00:17:10,566 --> 00:17:13,566
So imagine that 
you're a young child

380
00:17:13,666 --> 00:17:15,366
and you're learning 
about the category

381
00:17:15,466 --> 00:17:18,100
of bunny for the first time.

382
00:17:18,200 --> 00:17:20,466
You'll see one bunny 
in one context,

383
00:17:20,566 --> 00:17:22,733
and then you'll see another 
bunny in another context.

384
00:17:22,833 --> 00:17:26,366
Then you might see a third bunny 
in a different context,

385
00:17:26,466 --> 00:17:27,600
and what you need to do

386
00:17:27,700 --> 00:17:30,233
in order to learn 
the category of bunny

387
00:17:30,333 --> 00:17:33,266
is to abstract out the 
similarities and differences

388
00:17:33,366 --> 00:17:35,266
across these experiences.

389
00:17:35,366 --> 00:17:36,633
So you might see that,

390
00:17:36,733 --> 00:17:39,266
well, the bunnies 
share the same shape.

391
00:17:39,366 --> 00:17:41,666
And so maybe that's 
a relevant feature

392
00:17:41,766 --> 00:17:43,833
of the category of bunny.

393
00:17:43,933 --> 00:17:46,100
However, there are other 
features that vary.

394
00:17:46,200 --> 00:17:47,433
So for instance,

395
00:17:47,533 --> 00:17:50,566
the color of the hair is 
different across these bunnies.

396
00:17:50,666 --> 00:17:51,933
And a child might think,

397
00:17:52,033 --> 00:17:55,566
"Hmm, I guess the hair color 
is not a relevant category.

398
00:17:55,666 --> 00:17:58,166
It's irrelevant to the 
category of bunny."

399
00:17:59,166 --> 00:18:01,100
And the reason that 
children need to abstract

400
00:18:01,200 --> 00:18:02,300
across these experiences

401
00:18:02,400 --> 00:18:03,900
is they have to know 
what to generalize

402
00:18:04,000 --> 00:18:05,633
based on when they 
see new bunnies

403
00:18:05,733 --> 00:18:07,466
that they've never seen before.

404
00:18:08,466 --> 00:18:11,166
They should generalize 
based on body shape

405
00:18:11,266 --> 00:18:13,833
rather than on something 
like the color of the hair.

406
00:18:14,833 --> 00:18:17,633
Now, Ernie and others 
thought that the forgetting

407
00:18:17,733 --> 00:18:19,600
would deter this process

408
00:18:19,700 --> 00:18:21,433
because how are you 
supposed to abstract

409
00:18:21,533 --> 00:18:23,133
across all of your experiences

410
00:18:23,233 --> 00:18:26,000
if you can't remember 
your experiences?

411
00:18:26,100 --> 00:18:29,666
So for instance, if you can 
only remember the body shape

412
00:18:29,766 --> 00:18:31,466
of one bunny you've seen before,

413
00:18:31,566 --> 00:18:34,633
how are you supposed 
to know that body shape

414
00:18:34,733 --> 00:18:37,166
is a relevant feature 
to the category of bunny?

415
00:18:37,266 --> 00:18:39,566
You can't remember 
that all the bunnies

416
00:18:39,666 --> 00:18:41,766
shared the same body shape.

417
00:18:41,866 --> 00:18:44,433
Same thing goes 
for irrelevant features.

418
00:18:44,533 --> 00:18:46,533
If you only see 
one color of hair

419
00:18:46,633 --> 00:18:47,933
and you don't see that it varies

420
00:18:48,033 --> 00:18:49,733
across all the 
different bunnies,

421
00:18:49,833 --> 00:18:53,033
you might think that hair 
is a relevant feature.

422
00:18:53,133 --> 00:18:55,800
You know, the 
first bunny you see

423
00:18:55,900 --> 00:18:57,900
might be a certain color

424
00:18:58,000 --> 00:18:59,533
and you would think 
that that color

425
00:18:59,633 --> 00:19:03,233
defines the category of 
bunny when it does not.

426
00:19:03,333 --> 00:19:06,800
So in other words, forgetting 
should be bad for abstraction.

427
00:19:07,800 --> 00:19:10,400
Now what's interesting is that 
despite these fighting words

428
00:19:10,500 --> 00:19:13,333
being out there and strong-held 
belief by researchers,

429
00:19:13,433 --> 00:19:15,400
no one had really tested it.

430
00:19:15,500 --> 00:19:19,533
So what we decided to do 
was to test this hypothesis,

431
00:19:19,633 --> 00:19:22,733
and one of the most common 
methodologies that we've used

432
00:19:22,833 --> 00:19:24,100
to test this hypothesis

433
00:19:24,200 --> 00:19:27,566
is called the novel 
category induction task.

434
00:19:27,666 --> 00:19:29,866
And this task is designed 
for young children

435
00:19:29,966 --> 00:19:33,133
who are learning words and 
categories for the first time.

436
00:19:33,233 --> 00:19:36,300
And what we do is we present 
them with novel objects

437
00:19:36,400 --> 00:19:38,500
and novel words to ensure 
that they're learning

438
00:19:38,600 --> 00:19:41,566
these objects and words 
for the first time.

439
00:19:41,666 --> 00:19:43,666
So just as an example,

440
00:19:43,766 --> 00:19:45,900
an experimenter 
might show children

441
00:19:46,000 --> 00:19:48,933
a series of novel objects like 
this and label them and say,

442
00:19:49,033 --> 00:19:52,400
"This is a wug, this is a wug, 
and this is a wug."

443
00:19:52,500 --> 00:19:56,933
And then at a post-test, 
present children

444
00:19:57,033 --> 00:19:58,466
with a series of objects.

445
00:19:58,566 --> 00:20:01,033
Some they've never seen 
before, some they have.

446
00:20:01,133 --> 00:20:03,633
So for example, they'd 
seen a cat before.

447
00:20:03,733 --> 00:20:07,333
And then it's the child's job 
to generalize the category

448
00:20:07,433 --> 00:20:10,133
that they just learned to a 
new instance of the category.

449
00:20:10,233 --> 00:20:12,133
So what the experimenter 
would say is,

450
00:20:12,233 --> 00:20:13,833
"Can you hand me the wug?"

451
00:20:13,933 --> 00:20:17,366
And it would be the child's job 
to pick up that novel wug

452
00:20:17,466 --> 00:20:20,366
they haven't seen before that 
shares a similar feature,

453
00:20:20,466 --> 00:20:23,800
in this case shape, to the 
objects that they saw

454
00:20:23,900 --> 00:20:25,733
during learning.

455
00:20:25,833 --> 00:20:29,533
So this is referred to as the 
novel category induction task.

456
00:20:30,533 --> 00:20:32,800
And today I'm gonna show 
you two experiments,

457
00:20:32,900 --> 00:20:34,733
Experiments 1 and 2.

458
00:20:34,833 --> 00:20:37,300
And the participants in these 
experiments were typically

459
00:20:37,400 --> 00:20:39,500
developing two to two and a half 
year-old children.

460
00:20:39,600 --> 00:20:41,300
And again, we were 
really interested

461
00:20:41,400 --> 00:20:42,833
in how this process unfolds

462
00:20:42,933 --> 00:20:46,033
when it's organically unfolding 
a lot during development,

463
00:20:46,133 --> 00:20:48,166
which is the toddler period.

464
00:20:48,266 --> 00:20:50,966
They're learning lots of 
words and new categories

465
00:20:51,066 --> 00:20:52,566
during the toddler period.

466
00:20:53,566 --> 00:20:58,466
We made very simple manipulation 
on the classic paradigm.

467
00:20:58,566 --> 00:21:00,366
And what we did in Experiment 1

468
00:21:00,466 --> 00:21:04,666
is we presented children 
with novel words and objects

469
00:21:04,766 --> 00:21:09,433
on one of three schedules: 
simultaneous, massed, or spaced.

470
00:21:09,533 --> 00:21:10,600
I'm gonna walk you through

471
00:21:10,700 --> 00:21:13,200
each of these presentation 
schedules now.

472
00:21:13,300 --> 00:21:15,366
In the simultaneous condition,

473
00:21:15,466 --> 00:21:18,133
children were presented with 
four objects at the same time.

474
00:21:18,233 --> 00:21:21,466
They were all put on the 
table simultaneously.

475
00:21:21,566 --> 00:21:24,966
What the experimenter would do, 
would label each of the objects.

476
00:21:25,066 --> 00:21:26,300
So the experimenter would say,

477
00:21:26,400 --> 00:21:29,366
"Look, this is a wug," and 
point to the first object.

478
00:21:29,466 --> 00:21:30,533
They would wait 10 seconds

479
00:21:30,633 --> 00:21:32,333
and then point to the 
second object and say,

480
00:21:32,433 --> 00:21:34,300
"Look, this is a wug,"

481
00:21:34,400 --> 00:21:36,200
wait 10 seconds and 
so on and so forth

482
00:21:36,300 --> 00:21:38,833
for all four objects.

483
00:21:38,933 --> 00:21:41,800
Now in this condition, 
forgetting is minimized

484
00:21:41,900 --> 00:21:45,000
because all of the objects 
are present to the child

485
00:21:45,100 --> 00:21:46,300
at the same time.

486
00:21:46,400 --> 00:21:49,400
And they keep hearing the 
word wug over and over again.

487
00:21:49,500 --> 00:21:51,900
So children don't have 
to rely on their memory

488
00:21:52,000 --> 00:21:54,800
to abstract out what are the 
similarities and differences

489
00:21:54,900 --> 00:21:55,966
among the objects.

490
00:21:56,066 --> 00:21:58,966
They're all on the table 
at the same time.

491
00:22:00,400 --> 00:22:02,500
The second condition was 
the massed condition.

492
00:22:02,600 --> 00:22:03,733
And in this condition,

493
00:22:03,833 --> 00:22:06,700
we presented children 
with objects one at a time.

494
00:22:06,800 --> 00:22:09,466
So the experimenter would 
put an object on the table

495
00:22:09,566 --> 00:22:11,400
and say, "Look, this is a wug,"

496
00:22:11,500 --> 00:22:14,400
wait 10 seconds, take 
the object off the table,

497
00:22:14,500 --> 00:22:16,066
and then put it 
back on the table

498
00:22:16,166 --> 00:22:18,166
or put another object 
on the table and say,

499
00:22:18,266 --> 00:22:20,333
"Look, this is a wug,"

500
00:22:20,433 --> 00:22:22,533
wait 10 seconds for the child 
to play with the toy,

501
00:22:22,633 --> 00:22:24,900
take it off, and then 
put a third object on,

502
00:22:25,000 --> 00:22:26,333
and so on and so forth.

503
00:22:26,433 --> 00:22:29,666
So in other words, the objects 
were presented one at a time.

504
00:22:29,766 --> 00:22:31,200
Now in this case,

505
00:22:31,300 --> 00:22:34,166
children need to rely on 
their memory a little bit

506
00:22:34,266 --> 00:22:36,133
to remember what 
they've seen earlier.

507
00:22:36,233 --> 00:22:37,733
Because they only 
see one at a time,

508
00:22:37,833 --> 00:22:39,000
they have to think back,

509
00:22:39,100 --> 00:22:41,300
"Hmm, what were the colors 
and shapes of the objects

510
00:22:41,400 --> 00:22:42,766
that I saw earlier?"

511
00:22:44,100 --> 00:22:46,966
The final condition was 
the spaced condition.

512
00:22:47,066 --> 00:22:49,333
In this condition, 
what children did

513
00:22:49,433 --> 00:22:52,333
is they also saw one object 
at a time,

514
00:22:52,433 --> 00:22:54,833
but instead of seeing them 
in immediate succession,

515
00:22:54,933 --> 00:22:58,066
like in the massed condition, 
there were 30-second delays

516
00:22:58,166 --> 00:23:01,333
between each presentation 
of the object.

517
00:23:01,433 --> 00:23:04,033
So what children would do 
is they'd see an object,

518
00:23:04,133 --> 00:23:06,033
the experimenter would put 
one on the table and say,

519
00:23:06,133 --> 00:23:07,200
"Look, this is a wug."

520
00:23:07,300 --> 00:23:09,400
And then the experimenter 
would take it away.

521
00:23:09,500 --> 00:23:12,366
And there'd be 30 seconds 
of irrelevant activities

522
00:23:12,466 --> 00:23:15,600
like playing with Play-Doh, 
putting stickers on paper,

523
00:23:15,700 --> 00:23:19,000
anything to keep the kids 
at the table and entertained.

524
00:23:19,100 --> 00:23:20,933
After that 30-second 
play period,

525
00:23:21,033 --> 00:23:23,900
children would then be 
presented with another object

526
00:23:24,000 --> 00:23:26,833
and the experimenter would say, 
"Look, this is a wug."

527
00:23:26,933 --> 00:23:28,633
And that would happen 
so on and so forth

528
00:23:28,733 --> 00:23:30,933
for four objects in total.

529
00:23:32,633 --> 00:23:35,266
Now what you can see 
across these conditions

530
00:23:35,366 --> 00:23:37,766
is that the only thing 
that's being manipulated here

531
00:23:37,866 --> 00:23:38,933
is the timing.

532
00:23:39,033 --> 00:23:42,833
Either children see all objects 
simultaneously,

533
00:23:42,933 --> 00:23:45,533
preventing forgetting, or 
they see them one at a time.

534
00:23:45,633 --> 00:23:47,066
And what the spaced 
condition does

535
00:23:47,166 --> 00:23:49,333
is it introduces the most 
forgetting during learning

536
00:23:49,433 --> 00:23:51,833
because there are those 
30-second intervals

537
00:23:51,933 --> 00:23:53,733
between each presentation

538
00:23:53,833 --> 00:23:56,466
that gives children the 
opportunity to forget

539
00:23:56,566 --> 00:23:58,633
in between each object.

540
00:23:58,733 --> 00:24:00,966
And so they have to 
think harder and harder

541
00:24:01,066 --> 00:24:03,766
about what they saw earlier 
in order to abstract

542
00:24:03,866 --> 00:24:07,200
the relevant and irrelevant 
features of the category wug.

543
00:24:09,000 --> 00:24:12,833
Children were also presented 
with a distractor item

544
00:24:12,933 --> 00:24:16,033
during learning, and the 
experimenter did not label this.

545
00:24:16,133 --> 00:24:17,433
They would say, 
"Look at this toy."

546
00:24:17,533 --> 00:24:20,100
And it was presented for the 
same amount of time

547
00:24:20,200 --> 00:24:23,400
as the objects, 40 seconds.

548
00:24:23,500 --> 00:24:27,100
And then there was 
either a immediate test

549
00:24:27,200 --> 00:24:30,800
or a test after a delay, 
a 15-minute delay.

550
00:24:30,900 --> 00:24:33,933
And the reason for this 
difference across conditions

551
00:24:34,033 --> 00:24:35,133
is that we were interested

552
00:24:35,233 --> 00:24:37,233
in not only forgetting 
during learning,

553
00:24:37,333 --> 00:24:39,200
that's those three conditions 
I showed you earlier,

554
00:24:39,300 --> 00:24:41,166
but we were also 
interested in forgetting

555
00:24:41,266 --> 00:24:43,300
between learning and test.

556
00:24:43,400 --> 00:24:46,966
And so immediate condition 
prevents forgetting

557
00:24:47,066 --> 00:24:48,733
from happening 
between learning and test,

558
00:24:48,833 --> 00:24:52,466
and a 15-minute delay introduces 
the opportunity of forgetting

559
00:24:52,566 --> 00:24:56,900
between the immediate 
and delayed test.

560
00:24:57,000 --> 00:24:59,066
So we're introducing forgetting

561
00:24:59,166 --> 00:25:02,133
at different levels of 
the learning process.

562
00:25:02,233 --> 00:25:04,100
And then the test item 
was very similar

563
00:25:04,200 --> 00:25:05,733
to what I showed you earlier.

564
00:25:05,833 --> 00:25:08,633
Children would be presented 
with four objects,

565
00:25:08,733 --> 00:25:11,433
and then it was the 
child's job to pick out

566
00:25:11,533 --> 00:25:13,900
a novel instance of the category 
that they learned about.

567
00:25:14,000 --> 00:25:16,733
So the experimenter would say, 
"Can you hand me the wug?"

568
00:25:16,833 --> 00:25:18,800
And it's the child's job 
to pick up the wug

569
00:25:18,900 --> 00:25:21,800
and place it into the 
experimenter's hand.

570
00:25:23,733 --> 00:25:27,866
Okay, what I'm going to do 
is show you some results.

571
00:25:28,866 --> 00:25:30,500
This is the results 
of Experiment 1.

572
00:25:30,600 --> 00:25:34,333
And what you're seeing on the 
X-axis is the testing delay.

573
00:25:34,433 --> 00:25:37,100
So either children being 
tested immediately,

574
00:25:37,200 --> 00:25:40,066
or a 15-minute delay,

575
00:25:40,166 --> 00:25:42,933
and then on the Y-axis is the 
number of correct responses.

576
00:25:43,033 --> 00:25:45,066
So this is the mean number 
of times

577
00:25:45,166 --> 00:25:49,533
that children correctly handed 
the wug to the experimenter.

578
00:25:51,033 --> 00:25:53,833
So here's what we found 
at the immediate test.

579
00:25:53,933 --> 00:25:55,933
Children in the 
simultaneous condition

580
00:25:56,033 --> 00:25:58,166
significantly 
outperformed children

581
00:25:58,266 --> 00:26:00,700
in the massed and the 
spaced conditions.

582
00:26:01,700 --> 00:26:03,566
So if Ernie Rothkopf 
were here, he would say,

583
00:26:03,666 --> 00:26:05,066
"Look, I'm right.

584
00:26:05,166 --> 00:26:07,766
"The condition that had the 
least amount of forgetting

585
00:26:07,866 --> 00:26:09,733
led to the most learning."

586
00:26:09,833 --> 00:26:12,333
'Cause the simultaneous 
condition at the immediate test

587
00:26:12,433 --> 00:26:15,800
introduced no learning or no 
forgetting during learning,

588
00:26:15,900 --> 00:26:18,500
but also no forgetting 
between learning and test.

589
00:26:18,600 --> 00:26:21,133
So this would sort of prove 
Ernie and others right

590
00:26:21,233 --> 00:26:25,500
that forgetting does deter 
conceptual development.

591
00:26:25,600 --> 00:26:30,600
However, let's look at the data 
from the 15-minute delay.

592
00:26:31,800 --> 00:26:33,400
What we observed at the 
15-minute delay

593
00:26:33,500 --> 00:26:35,400
is that children in the 
spaced condition

594
00:26:35,500 --> 00:26:37,300
significantly 
outperformed children

595
00:26:37,400 --> 00:26:39,733
in the simultaneous 
and massed conditions.

596
00:26:40,733 --> 00:26:41,866
And you might wonder here,

597
00:26:41,966 --> 00:26:43,633
"Well, wow, this 
is an interaction,

598
00:26:43,733 --> 00:26:45,700
"a completely different 
pattern of results

599
00:26:45,800 --> 00:26:47,566
of the delayed test."

600
00:26:47,666 --> 00:26:49,166
And indeed, what we see here

601
00:26:49,266 --> 00:26:52,666
is children that had the most 
opportunities of forgetting

602
00:26:52,766 --> 00:26:56,200
perform the highest, because 
indeed the 15-minute delay

603
00:26:56,300 --> 00:26:57,433
was the most amount 
of forgetting

604
00:26:57,533 --> 00:26:59,000
between learning and test,

605
00:26:59,100 --> 00:27:01,566
and then the spaced condition 
introduced forgetting

606
00:27:01,666 --> 00:27:03,900
during the learning process.

607
00:27:04,000 --> 00:27:08,000
So this is the exact opposite 
of what Ernie would predict.

608
00:27:08,100 --> 00:27:10,233
Here what we see is 
that forgetting might be

609
00:27:10,333 --> 00:27:13,633
facilitating abstraction 
and generalization.

610
00:27:15,133 --> 00:27:19,166
I'm going to jump to 
Experiment 2 quickly

611
00:27:19,266 --> 00:27:22,266
and just explain why 
we did Experiment 2.

612
00:27:22,366 --> 00:27:23,433
So in Experiment 2,

613
00:27:23,533 --> 00:27:25,466
what we were interested 
in is we said,

614
00:27:25,566 --> 00:27:27,666
"Well, we think that 
forgetting is happening

615
00:27:27,766 --> 00:27:29,433
between learning events."

616
00:27:29,533 --> 00:27:32,866
In other words, we think 
that these 30-second gaps

617
00:27:32,966 --> 00:27:35,866
introduced opportunities 
for children to forget.

618
00:27:35,966 --> 00:27:38,266
But what we wanted to know 
is we wanted to have data

619
00:27:38,366 --> 00:27:41,166
to actually show that the 
forgetting is happening.

620
00:27:41,266 --> 00:27:43,433
So what we decided 
to do in Experiment 2

621
00:27:43,533 --> 00:27:48,300
is ask children to retrieve 
their learning during learning.

622
00:27:48,400 --> 00:27:49,800
And here's how we did this.

623
00:27:49,900 --> 00:27:52,000
So imagine children are in the 
simultaneous condition,

624
00:27:52,100 --> 00:27:55,766
they have four objects put on 
the table at the same time.

625
00:27:55,866 --> 00:27:58,366
And the experimenter would 
label the first object and say,

626
00:27:58,466 --> 00:28:00,166
"Look, this is a wug."

627
00:28:00,266 --> 00:28:03,266
Wait 10 seconds and then point 
to the second object and say,

628
00:28:03,366 --> 00:28:04,533
"What is this called?"

629
00:28:05,533 --> 00:28:06,700
They would wait five seconds,

630
00:28:06,800 --> 00:28:08,733
and then regardless of 
what the children said,

631
00:28:08,833 --> 00:28:11,400
the experimenter would say, 
"This is a wug."

632
00:28:11,500 --> 00:28:13,200
And then they'd move 
on to the third object,

633
00:28:13,300 --> 00:28:15,866
and they point to it and they'd 
say, "What is this called?"

634
00:28:15,966 --> 00:28:17,100
They'd wait five seconds,

635
00:28:17,200 --> 00:28:18,966
and then regardless of 
what the child said,

636
00:28:19,066 --> 00:28:20,633
they'd say, "This is a wug."

637
00:28:20,733 --> 00:28:23,300
And they did that 
for all four objects.

638
00:28:23,400 --> 00:28:25,166
The same thing 
happened in the massed

639
00:28:25,266 --> 00:28:26,766
and the spaced condition.

640
00:28:26,866 --> 00:28:28,600
Children would hear, 
"This is a wug,"

641
00:28:28,700 --> 00:28:29,966
on the first presentation.

642
00:28:30,066 --> 00:28:31,633
And then on the 
subsequent presentations,

643
00:28:31,733 --> 00:28:33,900
the experimenter would 
first prompt the children

644
00:28:34,000 --> 00:28:36,833
to report the word by saying, 
"What is this called?"

645
00:28:36,933 --> 00:28:40,600
Wait five seconds and 
then say, "This is a wug."

646
00:28:40,700 --> 00:28:42,500
And what these retrievals did

647
00:28:42,600 --> 00:28:45,466
is it gives us a sense 
of how readily children

648
00:28:45,566 --> 00:28:47,866
are retrieving information 
during learning.

649
00:28:48,866 --> 00:28:49,933
And what we predicted

650
00:28:50,033 --> 00:28:51,666
is that children in the 
spaced learning condition

651
00:28:51,766 --> 00:28:54,133
would demonstrate 
the most forgetting

652
00:28:54,233 --> 00:28:56,333
because of those 
30-second intervals.

653
00:28:57,866 --> 00:29:01,433
Okay, here's some data 
comparing the final test

654
00:29:01,533 --> 00:29:04,400
from Experiment 1 
and Experiment 2.

655
00:29:04,500 --> 00:29:07,066
And what you can see is 
we observe the same pattern

656
00:29:07,166 --> 00:29:08,633
of results across experiments.

657
00:29:08,733 --> 00:29:11,233
The bars are a little bit 
higher in Experiment 2.

658
00:29:11,333 --> 00:29:12,666
This is to be expected.

659
00:29:12,766 --> 00:29:15,433
It's because children are 
getting more retrieval practice

660
00:29:15,533 --> 00:29:18,033
by us asking them, 
"What is this called?"

661
00:29:18,133 --> 00:29:19,733
And so they overall did better,

662
00:29:19,833 --> 00:29:22,333
basically a retrieval 
practice effect here,

663
00:29:22,433 --> 00:29:25,000
but we see the same pattern, 
that at the immediate test,

664
00:29:25,100 --> 00:29:27,500
children do better in the 
simultaneous condition,

665
00:29:27,600 --> 00:29:29,033
but at the delayed test,

666
00:29:29,133 --> 00:29:31,133
they do better in the 
spaced condition.

667
00:29:33,000 --> 00:29:36,633
Here, what you're going 
to see are three bars

668
00:29:36,733 --> 00:29:39,266
representing the overall 
retrieval successes

669
00:29:39,366 --> 00:29:42,166
in the simultaneous, massed, 
and spaced conditions.

670
00:29:42,266 --> 00:29:43,766
So that's what will be 
on the X-axis.

671
00:29:43,866 --> 00:29:45,266
And then on the Y-axis,

672
00:29:45,366 --> 00:29:48,366
what you'll see is the mean 
number of retrieval successes

673
00:29:48,466 --> 00:29:49,733
out of 24.

674
00:29:49,833 --> 00:29:52,033
So there were eight trials 
with three retrieval attempts

675
00:29:52,133 --> 00:29:54,966
on each, so that adds up to 24.

676
00:29:55,066 --> 00:29:56,566
Here's what we found.

677
00:29:56,666 --> 00:29:59,966
We found that children in 
the simultaneous condition

678
00:30:00,066 --> 00:30:01,966
had significantly 
higher performance

679
00:30:02,066 --> 00:30:04,133
than children in the 
massed condition.

680
00:30:04,233 --> 00:30:05,933
And we found that children 
in the massed condition

681
00:30:06,033 --> 00:30:07,600
had significantly 
higher performance

682
00:30:07,700 --> 00:30:10,066
than children in the 
spaced condition.

683
00:30:10,166 --> 00:30:12,033
And this is exactly 
what we predicted.

684
00:30:12,133 --> 00:30:13,600
Children in the 
simultaneous condition

685
00:30:13,700 --> 00:30:15,333
are having a lot easier time

686
00:30:15,433 --> 00:30:17,633
retrieving the label 
for the category

687
00:30:17,733 --> 00:30:20,466
because they're not experiencing 
the same degree of forgetting

688
00:30:20,566 --> 00:30:23,066
as children in the massed 
or spaced conditions.

689
00:30:23,166 --> 00:30:25,900
And indeed, children 
in the spaced condition

690
00:30:26,000 --> 00:30:28,433
are struggling the most.

691
00:30:28,533 --> 00:30:32,666
We took this same data 
and we plotted it by time.

692
00:30:32,766 --> 00:30:34,966
And just as a reminder,

693
00:30:35,066 --> 00:30:37,533
what we did is we 
had Retrieval 1

694
00:30:37,633 --> 00:30:39,900
be at the second learning event,

695
00:30:40,000 --> 00:30:42,466
Retrieval 2 was at the 
third learning event,

696
00:30:42,566 --> 00:30:44,800
and Retrieval 3 was at the 
fourth learning event,

697
00:30:44,900 --> 00:30:46,066
'cause on that first 
learning event,

698
00:30:46,166 --> 00:30:47,233
we labeled the object.

699
00:30:47,333 --> 00:30:48,433
We said, "This is a wug."

700
00:30:48,533 --> 00:30:50,233
And then asked them 
what it was called

701
00:30:50,333 --> 00:30:51,566
on subsequent learning event.

702
00:30:51,666 --> 00:30:53,566
So on the X-axis, what 
you're seeing here

703
00:30:53,666 --> 00:30:55,833
is retrieval attempt 1, 2, 3,

704
00:30:55,933 --> 00:30:59,600
which corresponds to 
learning 2, 3, and 4.

705
00:30:59,700 --> 00:31:03,400
And we wanted to know if 
the nature of retrieval

706
00:31:03,500 --> 00:31:06,500
changed across the 
learning phase.

707
00:31:06,600 --> 00:31:08,233
So what you're seeing here first

708
00:31:08,333 --> 00:31:10,933
is performance for the 
simultaneous condition.

709
00:31:11,033 --> 00:31:13,200
And what you see as a 
relatively flat line

710
00:31:13,300 --> 00:31:15,500
across the three 
retrieval tasks.

711
00:31:15,600 --> 00:31:17,433
And what this suggests 
is that children

712
00:31:17,533 --> 00:31:19,266
in the simultaneous condition

713
00:31:19,366 --> 00:31:21,700
are observing overall 
high performance

714
00:31:21,800 --> 00:31:24,766
that doesn't change much 
across the learning phase.

715
00:31:25,800 --> 00:31:27,366
The next line that you're seeing

716
00:31:27,466 --> 00:31:29,233
is performance in the 
massed condition.

717
00:31:29,333 --> 00:31:30,566
It's a similar pattern

718
00:31:30,666 --> 00:31:32,900
to children in the 
simultaneous condition,

719
00:31:33,000 --> 00:31:34,066
it's just lower.

720
00:31:34,166 --> 00:31:35,400
So it's a little bit harder,

721
00:31:35,500 --> 00:31:37,100
but there's no change 
during learning

722
00:31:37,200 --> 00:31:39,466
for children in the 
massed condition.

723
00:31:39,566 --> 00:31:42,166
However, we see something 
fundamentally different

724
00:31:42,266 --> 00:31:43,900
in the spaced condition.

725
00:31:44,000 --> 00:31:46,500
What we observe is that the 
first retrieval attempt,

726
00:31:46,600 --> 00:31:49,300
children really struggle 
to retrieve information.

727
00:31:49,400 --> 00:31:52,600
In fact, it's not significantly 
different than zero.

728
00:31:52,700 --> 00:31:54,033
So essentially, when children

729
00:31:54,133 --> 00:31:55,533
get to that second 
retrieval attempt,

730
00:31:55,633 --> 00:31:57,033
they kind of stare 
at the experimenter

731
00:31:57,133 --> 00:31:58,433
and they don't know what to say.

732
00:31:58,533 --> 00:32:01,766
They're unable to 
produce the word wug.

733
00:32:01,866 --> 00:32:05,533
However, across the learning 
phase, we observed performance.

734
00:32:05,633 --> 00:32:09,866
So they get better and better 
at retrieving the word wug

735
00:32:09,966 --> 00:32:11,633
during learning.

736
00:32:11,733 --> 00:32:14,633
So what this data does is 
it confirms our hypothesis

737
00:32:14,733 --> 00:32:17,166
that this paradigm 
does introduce

738
00:32:17,266 --> 00:32:19,966
different levels of 
forgetting during learning,

739
00:32:20,066 --> 00:32:22,666
but what this data also does 
is it provides evidence

740
00:32:22,766 --> 00:32:25,400
for a theoretical count 
that explains

741
00:32:25,500 --> 00:32:27,933
why forgetting is 
promoting learning.

742
00:32:28,033 --> 00:32:30,866
And what we have argued 
is that forgetting

743
00:32:30,966 --> 00:32:34,566
is a form of abstraction, 
forgetting acts as abstraction

744
00:32:34,666 --> 00:32:37,633
in order to promote 
conceptual development.

745
00:32:37,733 --> 00:32:40,233
So I wanna walk you 
through now how this works.

746
00:32:41,233 --> 00:32:42,466
So imagine that you're a child

747
00:32:42,566 --> 00:32:46,066
and you're seeing a wug or a 
new toy for the first time.

748
00:32:46,166 --> 00:32:47,266
Well, you're gonna see that wug

749
00:32:47,366 --> 00:32:48,733
and you're immediately 
gonna forget

750
00:32:48,833 --> 00:32:50,066
everything about that wug

751
00:32:50,166 --> 00:32:51,933
according to the 
curvilinear pattern

752
00:32:52,033 --> 00:32:53,533
that I showed you earlier.

753
00:32:53,633 --> 00:32:55,566
And here, what we see is 
that forgetting curve.

754
00:32:56,566 --> 00:32:57,733
So you'll see a wug,

755
00:32:57,833 --> 00:32:59,566
and a certain amount 
of time will go on,

756
00:32:59,666 --> 00:33:01,900
and you will then 
experience another wug

757
00:33:02,000 --> 00:33:05,333
or another new toy that 
shares the category label.

758
00:33:05,433 --> 00:33:08,866
And what will happen 
is that that new wug

759
00:33:08,966 --> 00:33:12,733
will prompt you to recollect 
information that is similar

760
00:33:12,833 --> 00:33:15,500
to the information 
that you saw earlier.

761
00:33:15,600 --> 00:33:16,800
So in this case,

762
00:33:16,900 --> 00:33:21,400
what is similar across these 
two items is their shape.

763
00:33:21,500 --> 00:33:24,066
And so what you'll do is 
you'll retrieve the shape

764
00:33:24,166 --> 00:33:26,966
that you saw earlier 
and you'll engage

765
00:33:27,066 --> 00:33:28,800
in cognitive effort doing so

766
00:33:28,900 --> 00:33:30,800
because you've 
forgotten about it.

767
00:33:30,900 --> 00:33:32,600
And so you have to 
dig through your mind

768
00:33:32,700 --> 00:33:35,766
to find that information, 
that past memory.

769
00:33:35,866 --> 00:33:37,600
And what that 
cognitive effort does

770
00:33:37,700 --> 00:33:41,800
is it slows the 
memory for shape,

771
00:33:41,900 --> 00:33:45,133
which in this category 
is the relevant feature.

772
00:33:45,233 --> 00:33:47,900
Now that second object 
won't cue you to recall

773
00:33:48,000 --> 00:33:49,166
other types of information,

774
00:33:49,266 --> 00:33:51,666
like the color or the 
texture of the object

775
00:33:51,766 --> 00:33:52,833
that you saw earlier,

776
00:33:52,933 --> 00:33:55,866
because there isn't a 
match across the items.

777
00:33:55,966 --> 00:33:58,300
And so what that means 
is that you'll continue

778
00:33:58,400 --> 00:34:00,833
to forget that 
irrelevant information,

779
00:34:00,933 --> 00:34:05,066
the texture and the color 
at the same rate as earlier,

780
00:34:05,166 --> 00:34:07,366
when you first saw 
the original wug.

781
00:34:08,366 --> 00:34:11,666
And what you can see from 
looking at this image is now

782
00:34:11,766 --> 00:34:13,133
there's different 
forgetting rates

783
00:34:13,233 --> 00:34:15,100
for information that's relevant

784
00:34:15,200 --> 00:34:17,833
and information 
that's irrelevant.

785
00:34:18,833 --> 00:34:20,333
Let's do another 
demonstration of this.

786
00:34:20,433 --> 00:34:23,800
So let's just say 
that more time goes on

787
00:34:23,900 --> 00:34:27,233
and a child then 
sees yet another wug.

788
00:34:28,233 --> 00:34:30,366
And what that third wug 
is gonna do

789
00:34:30,466 --> 00:34:32,800
is prompt children 
to recall information

790
00:34:32,900 --> 00:34:35,966
from the first and second wug 
that's similar to that item.

791
00:34:36,066 --> 00:34:38,966
And again, that similar 
feature is shape.

792
00:34:39,066 --> 00:34:40,200
And so what children will do

793
00:34:40,300 --> 00:34:42,500
is they'll engage 
in cognitive effort,

794
00:34:42,600 --> 00:34:44,766
searching for that 
shape information.

795
00:34:44,866 --> 00:34:47,433
And what that cognitive effort 
will translate into

796
00:34:47,533 --> 00:34:51,833
is a slowed forgetting 
for that feature shape.

797
00:34:51,933 --> 00:34:54,733
So it's again, slowing 
the forgetting.

798
00:34:54,833 --> 00:34:57,833
However, features 
like color and texture

799
00:34:57,933 --> 00:34:59,566
are continuing to be forgot,

800
00:34:59,666 --> 00:35:03,000
according to that 
original forgetting curve.

801
00:35:03,100 --> 00:35:07,033
So they're being forgotten 
at a much different rate

802
00:35:07,133 --> 00:35:08,833
than the relevant information.

803
00:35:09,833 --> 00:35:12,933
And what we hypothesize happens 
across early development

804
00:35:13,033 --> 00:35:15,633
or learning about something 
new for the first time

805
00:35:15,733 --> 00:35:18,800
is this process of 
where we continually

806
00:35:18,900 --> 00:35:22,366
re-retrieve relevant 
information in the world

807
00:35:22,466 --> 00:35:25,866
and then we don't re-retrieve 
irrelevant information

808
00:35:25,966 --> 00:35:27,200
in the world.

809
00:35:27,300 --> 00:35:30,533
And as this process happens 
over and over and over again,

810
00:35:30,633 --> 00:35:34,100
we end up with different 
forgetting curves

811
00:35:34,200 --> 00:35:36,133
for information that's relevant

812
00:35:36,233 --> 00:35:38,700
and information 
that's irrelevant.

813
00:35:38,800 --> 00:35:41,333
And so by the time 
children or adults

814
00:35:41,433 --> 00:35:46,300
need to generalize to 
something new in that category,

815
00:35:46,400 --> 00:35:49,733
they will more readily 
retrieve relevant information

816
00:35:49,833 --> 00:35:52,666
like shape, compared to 
irrelevant information

817
00:35:52,766 --> 00:35:54,566
like color and texture.

818
00:35:54,666 --> 00:35:57,133
And this will support 
generalization

819
00:35:57,233 --> 00:35:58,466
and conceptual development

820
00:35:58,566 --> 00:36:01,966
because they're re-retrieving 
what's important

821
00:36:02,066 --> 00:36:05,566
and then applying it 
to that new category.

822
00:36:05,666 --> 00:36:08,933
So in other words, it's 
helping children, adults,

823
00:36:09,033 --> 00:36:12,133
everyone to abstract what's 
relevant in the world,

824
00:36:12,233 --> 00:36:16,800
and therefore helping us to 
learn categories and concepts.

825
00:36:18,833 --> 00:36:20,200
Just as a quick summary.

826
00:36:21,200 --> 00:36:25,033
For a long time, going 
back to the 1800s,

827
00:36:25,133 --> 00:36:27,500
researchers and general 
intuition told us

828
00:36:27,600 --> 00:36:30,500
that forgetting should deter 
conceptual development.

829
00:36:30,600 --> 00:36:32,966
It should prevent us from 
abstracting what's relevant

830
00:36:33,066 --> 00:36:35,200
and irrelevant in the world.

831
00:36:35,300 --> 00:36:38,900
However, my lab's work 
has shown the opposite:

832
00:36:39,000 --> 00:36:42,000
that forgetting can actually 
act as a form of abstraction,

833
00:36:42,100 --> 00:36:45,000
speed up abstraction, 
and help us to learn

834
00:36:45,100 --> 00:36:47,866
new categories and concepts.

835
00:36:47,966 --> 00:36:50,600
And so while we've always agreed

836
00:36:50,700 --> 00:36:53,333
that forgetting can help 
learning and memory,

837
00:36:53,433 --> 00:36:55,700
we now know that it can 
help us think and develop

838
00:36:55,800 --> 00:36:58,033
our most higher order 
forms of learning,

839
00:36:58,133 --> 00:37:01,433
like learning a new language 
or learning complex concepts.

840
00:37:02,966 --> 00:37:04,433
So that's the silver lining,

841
00:37:04,533 --> 00:37:08,700
is that even though forgetting 
is inevitable and predictable

842
00:37:08,800 --> 00:37:11,066
and will prevent us from 
remembering information,

843
00:37:11,166 --> 00:37:12,466
it's also gonna help us.

844
00:37:12,566 --> 00:37:14,600
It's gonna help us 
remember what's important

845
00:37:14,700 --> 00:37:17,533
so that we can generalize 
across our experiences

846
00:37:17,633 --> 00:37:19,666
based on the 
important information.

847
00:37:21,633 --> 00:37:23,600
So I have a number of 
concluding thoughts

848
00:37:23,700 --> 00:37:25,733
that I'd like to share 
with you all.

849
00:37:25,833 --> 00:37:27,500
The first concluding thought

850
00:37:27,600 --> 00:37:30,600
is to share some of the 
exciting work that we're doing,

851
00:37:30,700 --> 00:37:33,633
trying to build off this 
basic research

852
00:37:33,733 --> 00:37:35,933
to design interventions 
for children.

853
00:37:36,933 --> 00:37:38,833
One of the things that 
we've done is we've thought,

854
00:37:38,933 --> 00:37:43,233
"Well, why don't we integrate 
forgetting into curriculum?"

855
00:37:43,333 --> 00:37:46,300
In other words, we should 
work with teachers and parents

856
00:37:46,400 --> 00:37:48,766
in order to get 
them to encourage

857
00:37:48,866 --> 00:37:50,800
their children to forget.

858
00:37:50,900 --> 00:37:53,400
Now, I know this sounds 
very counterintuitive

859
00:37:53,500 --> 00:37:56,333
because wow, teachers are 
trying to help kids remember,

860
00:37:56,433 --> 00:37:58,133
parents are trying to 
help kids remember.

861
00:37:58,233 --> 00:38:02,266
But remember, forgetting 
can help us to remember.

862
00:38:02,366 --> 00:38:04,400
So what we've been doing 
is we've been working

863
00:38:04,500 --> 00:38:06,533
with classrooms and teachers

864
00:38:06,633 --> 00:38:10,066
and helping teachers to 
integrate spaced learning

865
00:38:10,166 --> 00:38:12,600
into their curriculum.

866
00:38:12,700 --> 00:38:14,133
Here's a quick example.

867
00:38:14,233 --> 00:38:17,133
In a lot of early 
childhood curriculum,

868
00:38:17,233 --> 00:38:18,633
teachers use units.

869
00:38:18,733 --> 00:38:22,100
So they might have a lion month 
or a flower month,

870
00:38:22,200 --> 00:38:23,766
where they talk 
about the same thing

871
00:38:23,866 --> 00:38:28,233
in extensive conversation 
for a month and then move on.

872
00:38:28,333 --> 00:38:30,266
So in other words, it's 
an immediate succession.

873
00:38:30,366 --> 00:38:32,600
It's kind of like cramming.

874
00:38:32,700 --> 00:38:35,166
So instead, what we've done 
is we've encouraged teachers

875
00:38:35,266 --> 00:38:38,766
to distribute lessons 
out across time.

876
00:38:38,866 --> 00:38:40,633
We've done a series 
of experiments

877
00:38:40,733 --> 00:38:43,300
to provide empirical evidence 
for teachers

878
00:38:43,400 --> 00:38:45,566
that this is a good thing to do.

879
00:38:45,666 --> 00:38:48,166
So for example, in some 
of our studies,

880
00:38:48,266 --> 00:38:50,566
we come up with different 
learning schedules

881
00:38:50,666 --> 00:38:52,000
for science curriculum.

882
00:38:52,100 --> 00:38:54,666
And the example that 
I'm gonna show here

883
00:38:54,766 --> 00:38:58,400
is the context of children's 
science curriculum

884
00:38:58,500 --> 00:39:00,466
about food chains.

885
00:39:00,566 --> 00:39:03,066
And food chains 
are a perfect context

886
00:39:03,166 --> 00:39:06,300
because what teachers do is 
they introduce food chains

887
00:39:06,400 --> 00:39:07,866
across different biomes.

888
00:39:07,966 --> 00:39:10,200
So they'll introduce the 
food chain in the grasslands

889
00:39:10,300 --> 00:39:11,800
and then the desert.

890
00:39:11,900 --> 00:39:13,400
And then what they 
expect children to do

891
00:39:13,500 --> 00:39:16,933
is to abstract out across all 
these different biomes

892
00:39:17,033 --> 00:39:19,566
in order to generalize 
to a new biome,

893
00:39:19,666 --> 00:39:22,066
like the Arctic or the swamp.

894
00:39:23,533 --> 00:39:26,200
And so what we do is we work 
with classrooms and teachers

895
00:39:26,300 --> 00:39:29,400
and develop a schedule such 
as the schedule you see here,

896
00:39:29,500 --> 00:39:32,433
where there's one of three 
presentations options.

897
00:39:32,533 --> 00:39:36,533
There's a massed schedule, 
a clumped schedule,

898
00:39:36,633 --> 00:39:38,600
and a spaced schedule.

899
00:39:38,700 --> 00:39:40,600
What you'll see in the 
massed schedule

900
00:39:40,700 --> 00:39:42,366
is that lessons 
about food chains

901
00:39:42,466 --> 00:39:44,300
are presented 
in immediate succession,

902
00:39:44,400 --> 00:39:47,933
just like we do in the 
lab with the word lists.

903
00:39:48,033 --> 00:39:51,433
The spaced condition 
distributes all the lessons

904
00:39:51,533 --> 00:39:53,000
out across days of the week

905
00:39:53,100 --> 00:39:55,500
so there's time in 
between each lesson,

906
00:39:55,600 --> 00:39:57,466
which maps onto the 
spaced schedules

907
00:39:57,566 --> 00:39:59,166
that we use in the lab as well.

908
00:40:00,166 --> 00:40:02,366
In these studies, we use 
an intermediate condition

909
00:40:02,466 --> 00:40:03,700
called the clump condition,

910
00:40:03,800 --> 00:40:06,733
which introduces a combination 
of massing and spacing.

911
00:40:06,833 --> 00:40:08,866
So some of the 
lessons are presented

912
00:40:08,966 --> 00:40:10,200
in immediate succession,

913
00:40:10,300 --> 00:40:14,066
whereas other lessons are 
distributed across time.

914
00:40:14,166 --> 00:40:17,966
And we have each lesson be 
about the same amount of time,

915
00:40:18,066 --> 00:40:21,100
so that we control for time 
across these experiments.

916
00:40:21,200 --> 00:40:24,433
And we have a test for 
children that happens one week

917
00:40:24,533 --> 00:40:26,166
after their last lesson

918
00:40:26,266 --> 00:40:30,800
to understand what they remember 
from the varied conditions.

919
00:40:32,366 --> 00:40:34,666
And I'm gonnashow 
you some data here,

920
00:40:34,766 --> 00:40:36,966
where we had two 
different types of tests,

921
00:40:37,066 --> 00:40:41,266
a simple generalization 
or simple concepts.

922
00:40:41,366 --> 00:40:43,833
This is things like learning 
that bigger creatures

923
00:40:43,933 --> 00:40:46,366
typically eat smaller creatures.

924
00:40:46,466 --> 00:40:49,700
And then we also had 
some complex concepts.

925
00:40:49,800 --> 00:40:53,100
And the idea that we tested here 
was the idea of interdependence.

926
00:40:53,200 --> 00:40:54,433
So in other words,

927
00:40:54,533 --> 00:40:57,800
if something happens to a 
particular creature in a biome,

928
00:40:57,900 --> 00:41:00,066
it affects all the 
other creatures.

929
00:41:00,166 --> 00:41:04,433
So for example, if there are no 
more fish in the ocean to eat,

930
00:41:04,533 --> 00:41:06,266
that affects all the 
other creatures

931
00:41:06,366 --> 00:41:07,666
that live in the ocean.

932
00:41:07,766 --> 00:41:10,066
They will have less to eat 
because there are no more fish,

933
00:41:10,166 --> 00:41:13,766
especially creatures higher up 
on the food chain.

934
00:41:14,766 --> 00:41:16,466
So that's what you're 
gonna see on the X-axis.

935
00:41:16,566 --> 00:41:19,033
On the Y-axis, 
what you're gonna see

936
00:41:19,133 --> 00:41:24,033
is the difference between the 
pre-test and the post-test.

937
00:41:24,133 --> 00:41:28,066
So we gave the tests before 
learning and after learning.

938
00:41:28,166 --> 00:41:31,633
And what you'll see is for the 
simple generalization

939
00:41:31,733 --> 00:41:33,066
or the simple concept,

940
00:41:33,166 --> 00:41:35,166
we found that children 
in the spaced conditions

941
00:41:35,266 --> 00:41:37,966
significantly outperformed 
children in the massed

942
00:41:38,066 --> 00:41:39,766
and the clumped conditions.

943
00:41:39,866 --> 00:41:42,366
And this is what we expected 
from prior research.

944
00:41:42,466 --> 00:41:45,366
We expected to see 
this based off of

945
00:41:45,466 --> 00:41:48,100
the relatively 
simple categories,

946
00:41:48,200 --> 00:41:50,933
like the wugs that 
you saw earlier.

947
00:41:51,933 --> 00:41:54,166
In this work, it was actually 
our first time testing

948
00:41:54,266 --> 00:41:57,000
whether or not we would 
find this phenomenon

949
00:41:57,100 --> 00:42:01,733
in a more complex task, like 
learning about interdependence,

950
00:42:01,833 --> 00:42:03,433
but let me show you the results.

951
00:42:03,533 --> 00:42:04,766
Here they are.

952
00:42:04,866 --> 00:42:06,200
We found the same pattern.

953
00:42:06,300 --> 00:42:08,333
We found that children 
in the spaced condition

954
00:42:08,433 --> 00:42:10,866
significantly outperformed 
children in the clumped

955
00:42:10,966 --> 00:42:13,033
and massed conditions.

956
00:42:13,133 --> 00:42:15,366
So what these 
classroom-based studies do

957
00:42:15,466 --> 00:42:17,733
is they help teachers 
to understand

958
00:42:17,833 --> 00:42:19,800
how they can 
manipulate forgetting

959
00:42:19,900 --> 00:42:22,600
by using spaced learning 
in their curriculum

960
00:42:22,700 --> 00:42:25,066
and in their teaching practices.

961
00:42:25,166 --> 00:42:28,733
And indeed, it's been a real joy 
for us to work one-on-one

962
00:42:28,833 --> 00:42:32,333
with teachers in classrooms 
to help improve STEM learning

963
00:42:32,433 --> 00:42:34,933
by just making these 
simple changes.

964
00:42:35,033 --> 00:42:37,766
It's amazing how 
just small changes

965
00:42:37,866 --> 00:42:39,833
in the timing of the lessons

966
00:42:39,933 --> 00:42:41,866
can create more learning 
in the classroom.

967
00:42:41,966 --> 00:42:45,266
And indeed, this is a 
free, innovative way

968
00:42:45,366 --> 00:42:46,500
to improve education.

969
00:42:46,600 --> 00:42:49,466
And we're actively 
studying how to develop

970
00:42:49,566 --> 00:42:51,200
these types of interventions

971
00:42:51,300 --> 00:42:54,266
so that we can encourage 
forgetting in the classroom

972
00:42:54,366 --> 00:42:56,666
and encourage teachers 
to think about

973
00:42:56,766 --> 00:43:00,233
how they can encourage 
forgetting in their students.

974
00:43:01,666 --> 00:43:04,633
So this is just an 
active area of research

975
00:43:04,733 --> 00:43:08,033
and is, you know, 
to be determined

976
00:43:08,133 --> 00:43:10,933
how much more 
we can do with this.

977
00:43:11,033 --> 00:43:12,333
You know, in another 
year or two,

978
00:43:12,433 --> 00:43:15,800
please contact me and 
we'll have more information

979
00:43:15,900 --> 00:43:18,533
about how the curriculum 
interventions are going.

980
00:43:19,600 --> 00:43:20,700
Another concluding thought

981
00:43:20,800 --> 00:43:22,500
that I really wanted 
to share with you today

982
00:43:22,600 --> 00:43:24,900
is how researchers 
are actively studying

983
00:43:25,000 --> 00:43:27,066
forgetting in the current times,

984
00:43:27,166 --> 00:43:30,300
namely during the 
COVID-19 pandemic.

985
00:43:30,400 --> 00:43:32,700
And if you're like me, 
you may have noticed,

986
00:43:32,800 --> 00:43:35,000
particularly during 
the lockdown period,

987
00:43:35,100 --> 00:43:37,500
that your thinking 
and learning have changed

988
00:43:37,600 --> 00:43:40,233
during the pandemic 
and in particular,

989
00:43:40,333 --> 00:43:43,366
that you've forgotten 
faster than you normally do.

990
00:43:43,466 --> 00:43:45,000
So you might lose your keys.

991
00:43:45,100 --> 00:43:47,566
You might not know 
what you had for lunch.

992
00:43:47,666 --> 00:43:50,366
You're forgetting things that 
you normally wouldn't forget.

993
00:43:50,466 --> 00:43:53,000
And there's a number 
of reasons of why

994
00:43:53,100 --> 00:43:55,566
you were experiencing 
that faster forgetting.

995
00:43:55,666 --> 00:43:59,666
Here's a quick summary of what 
research has found.

996
00:43:59,766 --> 00:44:04,466
So first, we know that 
there are routine changes.

997
00:44:04,566 --> 00:44:05,800
You're not going 
into the office,

998
00:44:05,900 --> 00:44:07,566
you're not going to the gym.

999
00:44:07,666 --> 00:44:10,366
And the reason that 
that affects our memory

1000
00:44:10,466 --> 00:44:14,366
is that our routines 
serve as memory cues.

1001
00:44:14,466 --> 00:44:16,500
We learn where we 
always put our keys

1002
00:44:16,600 --> 00:44:19,066
when we go to the office, 
where we put our lunch,

1003
00:44:19,166 --> 00:44:22,000
who we talk to and have 
chit-chat with in the hallway.

1004
00:44:22,100 --> 00:44:24,600
And when we lose that routine, 
we lose the cues.

1005
00:44:24,700 --> 00:44:28,566
And so what that's doing is 
it's speeding up forgetting.

1006
00:44:28,666 --> 00:44:30,866
If we don't have the cues, 
we're not gonna remember,

1007
00:44:30,966 --> 00:44:32,066
and we're gonna 
continue to forget

1008
00:44:32,166 --> 00:44:34,133
according to that faster rate.

1009
00:44:35,800 --> 00:44:38,600
The second reason 
that we observe

1010
00:44:38,700 --> 00:44:40,566
faster forgetting 
during the pandemic

1011
00:44:40,666 --> 00:44:44,900
has to do with a lack 
of social interaction.

1012
00:44:45,000 --> 00:44:47,533
Social interaction is 
important for so many reasons.

1013
00:44:47,633 --> 00:44:49,500
And I want to highlight two now.

1014
00:44:50,500 --> 00:44:51,633
So during the pandemic,

1015
00:44:51,733 --> 00:44:54,366
you haven't spent as 
much time with people

1016
00:44:54,466 --> 00:44:57,100
and people also 
serve as memory cues.

1017
00:44:57,200 --> 00:44:59,233
They prompt you to 
talk about your past.

1018
00:44:59,333 --> 00:45:01,500
So just imagine that you're 
chatting with a friend

1019
00:45:01,600 --> 00:45:03,200
or a family member, 
they might say,

1020
00:45:03,300 --> 00:45:04,633
"So what'd you do last week?"

1021
00:45:04,733 --> 00:45:05,833
And you responded with,

1022
00:45:05,933 --> 00:45:09,033
"Well, I got to talk 
to some friends.

1023
00:45:09,133 --> 00:45:11,433
I spent some time, 
it was so much fun."

1024
00:45:11,533 --> 00:45:14,433
And what you're doing 
by having that conversation

1025
00:45:14,533 --> 00:45:16,933
is you're recollecting 
your past experiences

1026
00:45:17,033 --> 00:45:19,300
and sharing them with that 
other person.

1027
00:45:19,400 --> 00:45:22,533
And so when people aren't 
around to prompt you,

1028
00:45:22,633 --> 00:45:25,633
you're not engaging as much 
practice in retrieving the past.

1029
00:45:25,733 --> 00:45:28,500
And again, that contributes 
to how fast you forget.

1030
00:45:29,666 --> 00:45:32,433
The other reason that social 
partners are so important

1031
00:45:32,533 --> 00:45:35,533
for our memory is 
that they contribute

1032
00:45:35,633 --> 00:45:37,066
to our emotional well-being.

1033
00:45:37,166 --> 00:45:39,100
And we know from a long history 
of research

1034
00:45:39,200 --> 00:45:40,666
on emotions and memory

1035
00:45:40,766 --> 00:45:43,166
is that when we're 
experiencing more volatile,

1036
00:45:43,266 --> 00:45:45,200
emotional experiences,

1037
00:45:45,300 --> 00:45:47,033
and when we're sort 
of experiencing

1038
00:45:47,133 --> 00:45:48,666
more negative life events,

1039
00:45:48,766 --> 00:45:52,800
that we actually forget faster 
during those times.

1040
00:45:52,900 --> 00:45:55,033
And so for those of you 
who are really missing

1041
00:45:55,133 --> 00:45:59,166
being with other people, 
and, you know,

1042
00:45:59,266 --> 00:46:02,633
are just sort of having 
a case of the blues,

1043
00:46:02,733 --> 00:46:04,300
you're forgetting faster may be

1044
00:46:04,400 --> 00:46:06,633
because of your 
current emotional state

1045
00:46:06,733 --> 00:46:08,533
or the emotional state 
that you experienced,

1046
00:46:08,633 --> 00:46:10,533
particularly during the lockdown

1047
00:46:10,633 --> 00:46:15,033
when you had little to no 
social interaction with people.

1048
00:46:15,133 --> 00:46:18,333
So there are many reasons 
that our social partners

1049
00:46:18,433 --> 00:46:22,966
are particularly important 
for our memory.

1050
00:46:24,133 --> 00:46:28,233
The third reason that we think 
people are forgetting faster

1051
00:46:28,333 --> 00:46:32,533
during the pandemic is that 
we're engaging in more behaviors

1052
00:46:32,633 --> 00:46:35,200
that speed up forgetting.

1053
00:46:35,300 --> 00:46:38,800
And one of those behaviors 
is alcohol consumption.

1054
00:46:38,900 --> 00:46:40,833
So there's a lot 
of research showing

1055
00:46:40,933 --> 00:46:42,366
that during the last 18 months,

1056
00:46:42,466 --> 00:46:44,633
people have been 
drinking more alcohol

1057
00:46:44,733 --> 00:46:46,333
and engaging in more drug use

1058
00:46:46,433 --> 00:46:49,233
than they did prior 
to the pandemic.

1059
00:46:49,333 --> 00:46:51,666
And we know from research 
on alcohol and drugs

1060
00:46:51,766 --> 00:46:56,033
that they act as a 
memory destabilizer

1061
00:46:56,133 --> 00:46:59,266
and therefore cause 
faster forgetting.

1062
00:46:59,366 --> 00:47:02,600
So if you're engaging 
in an over-consumption

1063
00:47:02,700 --> 00:47:07,033
of these substances, that can 
lead to complete memory loss.

1064
00:47:07,133 --> 00:47:08,700
Not only are you 
forgetting faster,

1065
00:47:08,800 --> 00:47:12,733
but you might not have 
any access to memories

1066
00:47:13,733 --> 00:47:16,900
from those states 
of over consumption.

1067
00:47:17,000 --> 00:47:19,800
So certainly we think that 
a lot of the forgetting

1068
00:47:19,900 --> 00:47:22,100
that's happening 
because of the pandemic

1069
00:47:22,200 --> 00:47:25,400
is because people are choosing 
to engage in behaviors

1070
00:47:25,500 --> 00:47:28,133
that are negative, that they 
for memory at least,

1071
00:47:28,233 --> 00:47:30,066
that they didn't before.

1072
00:47:31,766 --> 00:47:35,866
The last reason is that many 
of us during the pandemic

1073
00:47:35,966 --> 00:47:38,266
had lifestyle changes.

1074
00:47:38,366 --> 00:47:41,333
And one of the 
big lifestyle changes

1075
00:47:41,433 --> 00:47:43,766
is that many of us 
exercised less.

1076
00:47:43,866 --> 00:47:47,633
And this is even despite the 
public service announcements

1077
00:47:47,733 --> 00:47:49,966
in the news encouraging people

1078
00:47:50,066 --> 00:47:52,000
to get out there 
and go on a walk,

1079
00:47:52,100 --> 00:47:54,500
be outside, exercise.

1080
00:47:54,600 --> 00:47:57,900
And one of the reasons for that 
public service announcement

1081
00:47:58,000 --> 00:48:00,366
or why it was in the news 
all the time

1082
00:48:00,466 --> 00:48:03,100
is that we know that 
exercise is so valuable

1083
00:48:03,200 --> 00:48:04,333
for so many things,

1084
00:48:04,433 --> 00:48:06,166
our physical health, 
our mental health,

1085
00:48:06,266 --> 00:48:09,633
and the rate at which 
we forget information.

1086
00:48:09,733 --> 00:48:12,266
We have a long sort of 
big body of work showing

1087
00:48:12,366 --> 00:48:16,966
that for those of us who engage 
in more cardiovascular exercise,

1088
00:48:17,066 --> 00:48:19,633
but exercise more generally too,

1089
00:48:19,733 --> 00:48:24,000
that we are more, 
we're more readily able

1090
00:48:24,100 --> 00:48:27,333
to retrieve important 
information when we need to.

1091
00:48:27,433 --> 00:48:30,133
So in other words, 
exercise is good

1092
00:48:30,233 --> 00:48:34,533
and supports our ability 
to retrieve information.

1093
00:48:34,633 --> 00:48:38,333
And so, because of that, we,

1094
00:48:38,433 --> 00:48:41,866
you know, are forgetting faster 
if we're exercising less.

1095
00:48:41,966 --> 00:48:45,066
So in brief, there's a lot 
of lifestyle changes

1096
00:48:45,166 --> 00:48:47,966
that we made that aren't for the 
benefit of our memories.

1097
00:48:48,066 --> 00:48:49,933
And that's yet another reason

1098
00:48:50,033 --> 00:48:52,500
why we might be 
forgetting faster.

1099
00:48:52,600 --> 00:48:54,300
And researchers are 
actively studying,

1100
00:48:54,400 --> 00:48:56,066
now that we're coming 
out of the pandemic,

1101
00:48:56,166 --> 00:48:58,400
how our forgetting is changing.

1102
00:48:58,500 --> 00:49:02,100
We expect that once we establish 
those new routines,

1103
00:49:02,200 --> 00:49:03,533
live a healthier lifestyle,

1104
00:49:03,633 --> 00:49:05,566
that our forgetting 
will slow and go back

1105
00:49:05,666 --> 00:49:08,300
to be what it was pre-pandemic.

1106
00:49:08,400 --> 00:49:11,833
But we're also concerned that 
a small group of individuals

1107
00:49:11,933 --> 00:49:14,100
might not make those changes,

1108
00:49:14,200 --> 00:49:17,066
might continue to experience 
negative effects of the pandemic

1109
00:49:17,166 --> 00:49:18,766
and thus have faster forgetting.

1110
00:49:18,866 --> 00:49:21,400
So researchers are actively 
trying to figure out

1111
00:49:21,500 --> 00:49:23,866
how we can identify 
those individuals

1112
00:49:23,966 --> 00:49:25,700
so that we can help them.

1113
00:49:27,433 --> 00:49:29,833
My final concluding thought 
is just to say

1114
00:49:29,933 --> 00:49:32,833
that we don't know 
why we forget,

1115
00:49:32,933 --> 00:49:35,966
and we are actively 
studying it now,

1116
00:49:36,066 --> 00:49:40,633
but we do know that forgetting 
is essential for our lives.

1117
00:49:41,766 --> 00:49:43,366
Forgetting can help us 
to remember.

1118
00:49:43,466 --> 00:49:45,133
It can help us 
to learn new things,

1119
00:49:45,233 --> 00:49:47,666
like new words and concepts.

1120
00:49:47,766 --> 00:49:51,333
But most importantly, 
it can define who we are.

1121
00:49:51,433 --> 00:49:53,866
If we think back 
to the pandemic,

1122
00:49:53,966 --> 00:49:55,033
there's a lot of things

1123
00:49:55,133 --> 00:49:57,666
that we probably want 
to forget, right?

1124
00:49:57,766 --> 00:49:59,633
We had a lot of 
negative experiences,

1125
00:49:59,733 --> 00:50:02,066
particularly during 
the lockdown period,

1126
00:50:02,166 --> 00:50:04,166
and that forgetting is good.

1127
00:50:04,266 --> 00:50:06,100
So even though you may 
have forgotten faster

1128
00:50:06,200 --> 00:50:08,133
during the pandemic and 
forgotten a lot of the things

1129
00:50:08,233 --> 00:50:09,366
that happened to you,

1130
00:50:09,466 --> 00:50:11,333
I would argue that 
that's a good thing.

1131
00:50:11,433 --> 00:50:13,133
What your mind and 
brain are doing

1132
00:50:13,233 --> 00:50:17,066
is they're abstracting out 
what's important to remember,

1133
00:50:17,166 --> 00:50:21,766
hopefully the good things 
and the parts of yourself

1134
00:50:21,866 --> 00:50:25,400
that were able to endure 
during the pandemic.

1135
00:50:25,500 --> 00:50:26,800
And I'm hoping for all of you,

1136
00:50:26,900 --> 00:50:28,833
that you have a 
stronger sense of self

1137
00:50:28,933 --> 00:50:33,366
and know more about how you 
can endure in tough situations.

1138
00:50:33,466 --> 00:50:35,466
And forgetting is 
gonna help you do that.

1139
00:50:35,566 --> 00:50:38,566
It's gonna help you to 
remember when you were strong.

1140
00:50:39,566 --> 00:50:43,300
So on that note, I just 
want to say forgetting,

1141
00:50:43,400 --> 00:50:45,600
although intuitively may 
seem like a bad thing,

1142
00:50:45,700 --> 00:50:46,866
is a great thing,

1143
00:50:46,966 --> 00:50:49,566
and it's helping you with all 
facets of your life.

1144
00:50:50,566 --> 00:50:52,466
Thank you so much for listening.

1145
00:50:52,566 --> 00:50:53,666
Please feel free to reach out

1146
00:50:53,766 --> 00:50:55,000
if you have questions 
about forgetting.

1147
00:50:55,100 --> 00:50:57,366
As I mentioned at the 
beginning of the talk,

1148
00:50:57,466 --> 00:50:59,400
I spend lots of time 
thinking and learning

1149
00:50:59,500 --> 00:51:00,566
about thinking and learning,

1150
00:51:00,666 --> 00:51:02,966
and I would love to do that 
with you as well.

1151
00:51:03,066 --> 00:51:04,233
Good evening.