ABOUT THE SPEAKER
Petter Johansson - Experimental psychologist
Petter Johansson and his research group study self-knowledge and attitude change using methods ranging from questionnaires to close-up card magic.

Why you should listen

Petter Johansson is an associate professor in cognitive science, and together with Lars Hall he runs the Choice Blindness Lab at Lund University in Sweden. 
 
The main theme of Johansson's research is self-knowledge: How much do we know about ourselves, and how do we come to acquire this knowledge? To study these questions, he and his collaborators have developed an experimental paradigm known as "choice blindness." The methodological twist in these experiments is to use magic tricks to manipulate the outcome of people's choices -- and then measure to what extent and in what ways people react to these changes. The general finding is that participants often fail to detect when they receive the opposite of their choice, and when asked to explain, they readily construct and confabulate answers motivating a choice they only believe they intended to make. The effect has been demonstrated in choice experiments on topics such as facial attractiveness, consumer choice and moral and political decision making.

More profile about the speaker
Petter Johansson | Speaker | TED.com
TEDxUppsalaUniversity

Petter Johansson: Do you really know why you do what you do?

Filmed:
1,423,138 views

Experimental psychologist Petter Johansson researches choice blindness -- a phenomenon where we convince ourselves that we're getting what we want, even when we're not. In an eye-opening talk, he shares experiments (designed in collaboration with magicians!) that aim to answer the question: Why do we do what we do? The findings have big implications for the nature of self-knowledge and how we react in the face of manipulation. You may not know yourself as well as you think you do.
- Experimental psychologist
Petter Johansson and his research group study self-knowledge and attitude change using methods ranging from questionnaires to close-up card magic. Full bio

Double-click the English transcript below to play the video.

00:12
So why do you think
the rich should pay more in taxes?
0
800
3560
00:16
Why did you buy the latest iPhone?
1
4400
2376
00:18
Why did you pick your current partner?
2
6800
2456
00:21
And why did so many people
vote for Donald Trump?
3
9280
3416
00:24
What were the reasons, why did they do it?
4
12720
2520
00:27
So we ask this kind
of question all the time,
5
15990
2106
00:30
and we expect to get an answer.
6
18120
1736
00:31
And when being asked,
we expect ourselves to know the answer,
7
19880
3136
00:35
to simply tell why we did as we did.
8
23040
2480
00:38
But do we really know why?
9
26440
1720
00:41
So when you say that you prefer
George Clooney to Tom Hanks,
10
29000
3456
00:44
due to his concern for the environment,
11
32479
2057
00:46
is that really true?
12
34560
1200
00:48
So you can be perfectly sincere
and genuinely believe
13
36560
2496
00:51
that this is the reason
that drives your choice,
14
39080
2936
00:54
but to me, it may still feel
like something is missing.
15
42040
2600
00:57
As it stands, due to
the nature of subjectivity,
16
45560
3176
01:00
it is actually very hard to ever prove
that people are wrong about themselves.
17
48760
4320
01:06
So I'm an experimental psychologist,
18
54600
2136
01:08
and this is the problem
we've been trying to solve in our lab.
19
56760
3536
01:12
So we wanted to create an experiment
20
60320
2176
01:14
that would allow us to challenge
what people say about themselves,
21
62520
3536
01:18
regardless of how certain they may seem.
22
66080
2680
01:21
But tricking people
about their own mind is hard.
23
69960
2736
01:24
So we turned to the professionals.
24
72720
2376
01:27
The magicians.
25
75120
1200
01:29
So they're experts at creating
the illusion of a free choice.
26
77120
2896
01:32
So when they say, "Pick a card, any card,"
27
80040
2296
01:34
the only thing you know
is that your choice is no longer free.
28
82360
2920
01:38
So we had a few fantastic
brainstorming sessions
29
86200
2376
01:40
with a group of Swedish magicians,
30
88600
1856
01:42
and they helped us create a method
31
90480
1642
01:44
in which we would be able to manipulate
the outcome of people's choices.
32
92147
3973
01:48
This way we would know
when people are wrong about themselves,
33
96760
2936
01:51
even if they don't know this themselves.
34
99720
2040
01:54
So I will now show you
a short movie showing this manipulation.
35
102480
4656
01:59
So it's quite simple.
36
107160
1416
02:00
The participants make a choice,
37
108600
2136
02:02
but I end up giving them the opposite.
38
110760
2256
02:05
And then we want to see:
How did they react, and what did they say?
39
113040
3520
02:09
So it's quite simple, but see
if you can spot the magic going on.
40
117240
3160
02:13
And this was shot with real participants,
they don't know what's going on.
41
121440
3520
02:19
(Video) Petter Johansson:
Hi, my name's Petter.
42
127000
2216
02:21
Woman: Hi, I'm Becka.
43
129240
1215
02:22
PJ: I'm going to show you
pictures like this.
44
130479
2137
02:24
And you'll have to decide
which one you find more attractive.
45
132640
2896
02:27
Becka: OK.
46
135560
1216
02:28
PJ: And then sometimes,
I will ask you why you prefer that face.
47
136800
3176
02:32
Becka: OK.
48
140000
1216
02:33
PJ: Ready?
Becka: Yeah.
49
141240
1200
02:43
PJ: Why did you prefer that one?
50
151120
1816
02:44
Becka: The smile, I think.
51
152960
1496
02:46
PJ: Smile.
52
154480
1200
02:52
Man: One on the left.
53
160400
1240
02:57
Again, this one just struck me.
54
165520
1640
02:59
Interesting shot.
55
167760
1616
03:01
Since I'm a photographer,
I like the way it's lit and looks.
56
169400
3000
03:06
Petter Johansson: But now comes the trick.
57
174280
2040
03:10
(Video) Woman 1: This one.
58
178120
1280
03:16
PJ: So they get the opposite
of their choice.
59
184240
2280
03:20
And let's see what happens.
60
188520
1600
03:28
Woman 2: Um ...
61
196240
1200
03:35
I think he seems a little more
innocent than the other guy.
62
203760
2800
03:45
Man: The one on the left.
63
213360
1240
03:49
I like her smile
and contour of the nose and face.
64
217280
3696
03:53
So it's a little more interesting
to me, and her haircut.
65
221000
2760
04:00
Woman 3: This one.
66
228040
1200
04:03
I like the smirky look better.
67
231520
1576
04:05
PJ: You like the smirky look better?
68
233120
2000
04:09
(Laughter)
69
237680
3176
04:12
Woman 3: This one.
70
240880
1200
04:15
PJ: What made you choose him?
71
243280
1400
04:17
Woman 3: I don't know,
he looks a little bit like the Hobbit.
72
245520
2896
04:20
(Laughter)
73
248440
2056
04:22
PJ: And what happens in the end
74
250520
1496
04:24
when I tell them the true nature
of the experiment?
75
252040
3096
04:27
Yeah, that's it. I just have to
ask a few questions.
76
255160
2456
04:29
Man: Sure.
77
257640
1216
04:30
PJ: What did you think
of this experiment, was it easy or hard?
78
258880
2976
04:33
Man: It was easy.
79
261880
1240
04:36
PJ: During the experiments,
80
264040
1336
04:37
I actually switched
the pictures three times.
81
265400
3336
04:40
Was this anything you noticed?
82
268760
1576
04:42
Man: No. I didn't notice any of that.
83
270360
1816
04:44
PJ: Not at all?
Man: No.
84
272200
1496
04:45
Switching the pictures as far as ...
85
273720
2096
04:47
PJ: Yeah, you were pointing at one of them
but I actually gave you the opposite.
86
275840
3816
04:51
Man: The opposite one.
OK, when you --
87
279680
1816
04:53
No. Shows you how much
my attention span was.
88
281520
2256
04:55
(Laughter)
89
283800
1520
04:58
PJ: Did you notice that sometimes
during the experiment
90
286880
3016
05:01
I switched the pictures?
91
289920
2136
05:04
Woman 2: No, I did not notice that.
92
292080
2016
05:06
PJ: You were pointing at one,
but then I gave you the other one.
93
294120
3000
05:09
No inclination of that happening?
94
297920
1616
05:11
Woman 2: No.
95
299560
1576
05:13
Woman 2: I did not notice.
96
301160
1256
05:14
(Laughs)
97
302440
1936
05:16
PJ: Thank you.
98
304400
1216
05:17
Woman 2: Thank you.
99
305640
1376
05:19
PJ: OK, so as you probably
figured out now,
100
307040
2056
05:21
the trick is that I have
two cards in each hand,
101
309120
2256
05:23
and when I hand one of them over,
102
311400
1576
05:25
the black one kind of disappears
into the black surface on the table.
103
313000
4360
05:30
So using pictures like this,
104
318640
1736
05:32
normally not more than 20 percent
of the participants detect these tries.
105
320400
4376
05:36
And as you saw in the movie,
106
324800
1416
05:38
when in the end
we explain what's going on,
107
326240
3176
05:41
they're very surprised and often refuse
to believe the trick has been made.
108
329440
4376
05:45
So this shows that this effect
is quite robust and a genuine effect.
109
333840
4776
05:50
But if you're interested
in self-knowledge, as I am,
110
338640
2656
05:53
the more interesting bit is,
111
341320
1336
05:54
OK, so what did they say
when they explained these choices?
112
342680
3936
05:58
So we've done a lot of analysis
113
346640
1496
06:00
of the verbal reports
in these experiments.
114
348160
2080
06:03
And this graph simply shows
115
351360
2456
06:05
that if you compare
what they say in a manipulated trial
116
353840
4776
06:10
with a nonmanipulated trial,
117
358640
1376
06:12
that is when they explain
a normal choice they've made
118
360040
2776
06:14
and one where we manipulated the outcome,
119
362840
2496
06:17
we find that they are remarkably similar.
120
365360
2456
06:19
So they are just as emotional,
just as specific,
121
367840
3056
06:22
and they are expressed
with the same level of certainty.
122
370920
3200
06:27
So the strong conclusion to draw from this
123
375120
2336
06:29
is that if there are no differences
124
377480
2216
06:31
between a real choice
and a manipulated choice,
125
379720
3696
06:35
perhaps we make things up all the time.
126
383440
2440
06:38
But we've also done studies
127
386680
1336
06:40
where we try to match what they say
with the actual faces.
128
388040
3016
06:43
And then we find things like this.
129
391080
1880
06:45
So here, this male participant,
he preferred the girl to the left,
130
393760
5056
06:50
he ended up with the one to the right.
131
398840
1856
06:52
And then, he explained
his choice like this.
132
400720
2816
06:55
"She is radiant.
133
403560
1296
06:56
I would rather have approached her
at the bar than the other one.
134
404880
3096
07:00
And I like earrings."
135
408000
1616
07:01
And whatever made him choose
the girl on the left to begin with,
136
409640
3496
07:05
it can't have been the earrings,
137
413160
1576
07:06
because they were actually
sitting on the girl on the right.
138
414760
2856
07:09
So this is a clear example
of a post hoc construction.
139
417640
3776
07:13
So they just explained
the choice afterwards.
140
421440
2800
07:17
So what this experiment shows is,
141
425320
2296
07:19
OK, so if we fail to detect
that our choices have been changed,
142
427640
3656
07:23
we will immediately start
to explain them in another way.
143
431320
3200
07:27
And what we also found
144
435520
1256
07:28
is that the participants
often come to prefer the alternative,
145
436800
3216
07:32
that they were led to believe they liked.
146
440040
2256
07:34
So if we let them do the choice again,
147
442320
2016
07:36
they will now choose the face
they had previously rejected.
148
444360
3760
07:41
So this is the effect
we call "choice blindness."
149
449520
2296
07:43
And we've done
a number of different studies --
150
451840
2216
07:46
we've tried consumer choices,
151
454080
2536
07:48
choices based on taste and smell
and even reasoning problems.
152
456640
4416
07:53
But what you all want to know is of course
153
461080
2056
07:55
does this extend also
to more complex, more meaningful choices?
154
463160
3936
07:59
Like those concerning
moral and political issues.
155
467120
3080
08:04
So the next experiment,
it needs a little bit of a background.
156
472400
4216
08:08
So in Sweden, the political landscape
157
476640
4256
08:12
is dominated by a left-wing
and a right-wing coalition.
158
480920
3360
08:17
And the voters may move a little bit
between the parties within each coalition,
159
485720
4416
08:22
but there is very little movement
between the coalitions.
160
490160
2760
08:25
And before each elections,
161
493680
1976
08:27
the newspapers and the polling institutes
162
495680
4216
08:31
put together what they call
"an election compass"
163
499920
2616
08:34
which consists of a number
of dividing issues
164
502560
3336
08:37
that sort of separates the two coalitions.
165
505920
2336
08:40
Things like if tax on gasoline
should be increased
166
508280
3735
08:44
or if the 13 months of paid parental leave
167
512039
4096
08:48
should be split equally
between the two parents
168
516159
2496
08:50
in order to increase gender equality.
169
518679
2721
08:54
So, before the last Swedish election,
170
522840
2216
08:57
we created an election compass of our own.
171
525080
2600
09:00
So we walked up to people in the street
172
528480
2136
09:02
and asked if they wanted
to do a quick political survey.
173
530640
3336
09:06
So first we had them state
their voting intention
174
534000
2456
09:08
between the two coalitions.
175
536480
1360
09:10
Then we asked them
to answer 12 of these questions.
176
538560
3776
09:14
They would fill in their answers,
177
542360
1976
09:16
and we would ask them to discuss,
178
544360
1616
09:18
so OK, why do you think
tax on gas should be increased?
179
546000
5496
09:23
And we'd go through the questions.
180
551520
2096
09:25
Then we had a color coded template
181
553640
3896
09:29
that would allow us
to tally their overall score.
182
557560
2936
09:32
So this person would have
one, two, three, four
183
560520
3456
09:36
five, six, seven, eight, nine
scores to the left,
184
564000
3296
09:39
so he would lean to the left, basically.
185
567320
2680
09:42
And in the end, we also had them
fill in their voting intention once more.
186
570800
4440
09:48
But of course, there was
also a trick involved.
187
576160
2280
09:51
So first, we walked up to people,
188
579360
2176
09:53
we asked them
about their voting intention
189
581560
2056
09:55
and then when they started filling in,
190
583640
2256
09:57
we would fill in a set of answers
going in the opposite direction.
191
585920
5456
10:03
We would put it under the notepad.
192
591400
2576
10:06
And when we get the questionnaire,
193
594000
2776
10:08
we would simply glue it on top
of the participant's own answer.
194
596800
3320
10:16
So there, it's gone.
195
604000
1240
10:24
And then we would ask
about each of the questions:
196
612280
2376
10:26
How did you reason here?
197
614680
1536
10:28
And they'll state the reasons,
198
616240
1736
10:30
together we will sum up
their overall score.
199
618000
2480
10:34
And in the end, they will state
their voting intention again.
200
622800
3680
10:41
So what we find first of all here,
201
629960
1656
10:43
is that very few of these
manipulations are detected.
202
631640
4216
10:47
And they're not detected
in the sense that they realize,
203
635880
2656
10:50
"OK, you must have changed my answer,"
204
638560
1856
10:52
it was more the case that,
205
640440
1256
10:53
"OK, I must've misunderstood
the question the first time I read it.
206
641720
3176
10:56
Can I please change it?"
207
644920
1240
10:59
And even if a few of these
manipulations were changed,
208
647080
5136
11:04
the overall majority was missed.
209
652240
2136
11:06
So we managed to switch 90 percent
of the participants' answers
210
654400
3656
11:10
from left to right, right to left,
their overall profile.
211
658080
3160
11:14
And what happens then when
they are asked to motivate their choices?
212
662800
4400
11:20
And here we find much more
interesting verbal reports
213
668160
3056
11:23
than compared to the faces.
214
671240
2016
11:25
People say things like this,
and I'll read it to you.
215
673280
3360
11:29
So, "Large-scale governmental surveillance
of email and internet traffic
216
677720
3736
11:33
ought to be permissible as means to combat
international crime and terrorism."
217
681480
4336
11:37
"So you agree to some extent
with this statement." "Yes."
218
685840
2716
11:40
"So how did you reason here?"
219
688580
1500
11:43
"Well, like, as it is so hard to get
at international crime and terrorism,
220
691600
4936
11:48
I think there should be
those kinds of tools."
221
696560
2776
11:51
And then the person remembers an argument
from the newspaper in the morning.
222
699360
3616
11:55
"Like in the newspaper today,
223
703000
1616
11:56
it said they can like,
listen to mobile phones from prison,
224
704640
3376
12:00
if a gang leader tries to continue
his crimes from inside.
225
708040
3536
12:03
And I think it's madness
that we have so little power
226
711600
2816
12:06
that we can't stop those things
227
714440
1656
12:08
when we actually have
the possibility to do so."
228
716120
2936
12:11
And then there's a little bit
back and forth in the end:
229
719080
2696
12:13
"I don't like that they have access
to everything I do,
230
721800
2576
12:16
but I still think
it's worth it in the long run."
231
724400
2576
12:19
So, if you didn't know that this person
232
727000
2536
12:21
just took part in
a choice blindness experiment,
233
729560
2256
12:23
I don't think you would question
234
731840
1856
12:25
that this is the true attitude
of that person.
235
733720
3120
12:29
And what happens in the end,
with the voting intention?
236
737800
2856
12:32
What we find -- that one is also
clearly affected by the questionnaire.
237
740680
4696
12:37
So we have 10 participants
238
745400
1736
12:39
shifting from left to right
or from right to left.
239
747160
2976
12:42
We have another 19
that go from clear voting intention
240
750160
2536
12:44
to being uncertain.
241
752720
1456
12:46
Some go from being uncertain
to clear voting intention.
242
754200
3096
12:49
And then there is a number of participants
staying uncertain throughout.
243
757320
4736
12:54
And that number is interesting
244
762080
1576
12:55
because if you look
at what the polling institutes say
245
763680
4616
13:00
the closer you get to an election,
246
768320
1656
13:02
the only people that are sort of in play
247
770000
2136
13:04
are the ones that are
considered uncertain.
248
772160
2656
13:06
But we show there is a much larger number
249
774840
3216
13:10
that would actually consider
shifting their attitudes.
250
778080
2800
13:13
And here I must point out, of course,
that you are not allowed to use this
251
781640
3496
13:17
as an actual method
to change people's votes
252
785160
2616
13:19
before an election,
253
787800
1496
13:21
and we clearly debriefed them afterwards
254
789320
3616
13:24
and gave them every
opportunity to change back
255
792960
2296
13:27
to whatever they thought first.
256
795280
2480
13:30
But what this shows is
that if you can get people
257
798600
2336
13:32
to see the opposite view and engage
in a conversation with themselves,
258
800960
5536
13:38
that could actually make them
change their views.
259
806520
2920
13:42
OK.
260
810400
1200
13:44
So what does it all mean?
261
812760
1656
13:46
What do I think is going on here?
262
814440
2416
13:48
So first of all,
263
816880
1216
13:50
a lot of what we call self-knowledge
is actually self-interpretation.
264
818120
4856
13:55
So I see myself make a choice,
265
823000
2496
13:57
and then when I'm asked why,
266
825520
2776
14:00
I just try to make
as much sense of it as possible
267
828320
2536
14:02
when I make an explanation.
268
830880
1936
14:04
But we do this so quickly
and with such ease
269
832840
3016
14:07
that we think we actually know the answer
when we answer why.
270
835880
4280
14:13
And as it is an interpretation,
271
841040
3096
14:16
of course we sometimes make mistakes.
272
844160
2296
14:18
The same way we make mistakes
when we try to understand other people.
273
846480
3520
14:23
So beware when you ask people
the question "why"
274
851160
3696
14:26
because what may happen
is that, if you asked them,
275
854880
4896
14:31
"So why do you support this issue?"
276
859800
4016
14:35
"Why do you stay in this job
or this relationship?" --
277
863840
3216
14:39
what may happen when you ask why
is that you actually create an attitude
278
867080
3416
14:42
that wasn't there
before you asked the question.
279
870520
2240
14:45
And this is of course important
in your professional life, as well,
280
873440
3176
14:48
or it could be.
281
876640
1216
14:49
If, say, you design something
and then you ask people,
282
877880
2536
14:52
"Why do you think this is good or bad?"
283
880440
2256
14:54
Or if you're a journalist
asking a politician,
284
882720
3056
14:57
"So, why did you make this decision?"
285
885800
2376
15:00
Or if indeed you are a politician
286
888200
1936
15:02
and try to explain
why a certain decision was made.
287
890160
2640
15:06
So this may all seem a bit disturbing.
288
894080
3576
15:09
But if you want to look at it
from a positive direction,
289
897680
3496
15:13
it could be seen as showing,
290
901200
1736
15:14
OK, so we're actually
a little bit more flexible than we think.
291
902960
3376
15:18
We can change our minds.
292
906360
1896
15:20
Our attitudes are not set in stone.
293
908280
2456
15:22
And we can also change
the minds of others,
294
910760
3176
15:25
if we can only get them
to engage with the issue
295
913960
2376
15:28
and see it from the opposite view.
296
916360
1680
15:31
And in my own personal life,
since starting with this research --
297
919400
3936
15:35
So my partner and I,
we've always had the rule
298
923360
2576
15:37
that you're allowed to take things back.
299
925960
2296
15:40
Just because I said
I liked something a year ago,
300
928280
2336
15:42
doesn't mean I have to like it still.
301
930640
2040
15:45
And getting rid of the need
to stay consistent
302
933480
2816
15:48
is actually a huge relief and makes
relational life so mush easier to live.
303
936320
4360
15:53
Anyway, so the conclusion must be:
304
941720
2360
15:57
know that you don't know yourself.
305
945320
2496
15:59
Or at least not as well
as you think you do.
306
947840
2320
16:03
Thanks.
307
951480
1216
16:04
(Applause)
308
952720
4640

▲Back to top

ABOUT THE SPEAKER
Petter Johansson - Experimental psychologist
Petter Johansson and his research group study self-knowledge and attitude change using methods ranging from questionnaires to close-up card magic.

Why you should listen

Petter Johansson is an associate professor in cognitive science, and together with Lars Hall he runs the Choice Blindness Lab at Lund University in Sweden. 
 
The main theme of Johansson's research is self-knowledge: How much do we know about ourselves, and how do we come to acquire this knowledge? To study these questions, he and his collaborators have developed an experimental paradigm known as "choice blindness." The methodological twist in these experiments is to use magic tricks to manipulate the outcome of people's choices -- and then measure to what extent and in what ways people react to these changes. The general finding is that participants often fail to detect when they receive the opposite of their choice, and when asked to explain, they readily construct and confabulate answers motivating a choice they only believe they intended to make. The effect has been demonstrated in choice experiments on topics such as facial attractiveness, consumer choice and moral and political decision making.

More profile about the speaker
Petter Johansson | Speaker | TED.com