[MUSIC PLAYING]
BLAKE MCKIMMIE: Well, Bill, thanks for joining us today. BILL VON HIPPEL: My pleasure.
BLAKE MCKIMMIE: So, Bill, we've been talking a bit about deception and people's ability
to detect deception.
And one of the challenges that people face is they often think that they're very good
at detecting deception.
But the research suggests that they're not.
And one of the reasons is because people have a stereotype that's a bit inaccurate about
what nonverbal behavior should look like.
When somebody is being deceptive, they focus on all of these things.
But they're just indicators of nervousness.
And so this has kind of led people to conclude that people just generally are not very accurate
at deception detection.
Is that really the conclusion we should draw from their research?
BILL VON HIPPEL: Well, first of all I definitely agree with you that people make the mistake
of equating nervousness with deceiving.
And, of course, if you're interrogating me and I'm worried what happens if you don't
believe me, I'll be nervous.
Even if I know I'm completely innocent.
So I think those data are all very reliable.
But I think it's too big of an inference to then go on and say people can't detect deception.
Now, I'm making this argument in the face of a huge database that shows that people
can't do it.
So, of course, that raises the question, why would you make such an argument?
And I think the answer to that question is that the research has been done in a way that
facilitated ease of research, but didn't facilitate matching real world deception.
So real world deceptions are typically important.
I'm not just lying to you about your haircut I'm lying to you about something that, if
I'm discovered, I'll suffer for it.
And they're often quite complicated.
We eliminate all that in our typical research paradigm, where we just videotape somebody
who's telling a simple lie, and then I watch a videotape and I try to say if they're telling
the truth or not.
It's great for experimental control.
It's lousy for getting at the underlying dimensions of what really is involved when people deceive.
BLAKE MCKIMMIE: So there's are no real consequences in the lab for getting it wrong, like when
you're trying to figure out if someone's lying.
And for the lair themselves on the video, what have they got vested in trying to dupe
someone?
BILL VON HIPPEL: That's right.
They don't care if they're caught.
And the lie the telling is typically very simple.
That they're asked to go in the professor's office, steal the wallet off the table, or
don't.
And then they have to claim they didn't.
And we don't know which it is that they really did.
So that's a super-easy lie.
I didn't take the wallet, right?
You just keep saying that over and over again.
And you're on videotape, and you're not going to suffer if people believe you did.
But the real world, we often tell very complicated lies.
They involve reshaping all sorts of world events.
So that I really wasn't at the bowling alley with your wife at that time, or whatever the
case might be.
And it's very easy to get caught out in those.
And so, in our lab, we've done a little bit of work on this, where we try to get people
to engage in some very complicated lies.
And then see if people can detect them.
And what we found today is that, in fact, when people tell these complicated lies, there's
a bit of a truth bias.
And people believe that they're true.
They just automatically accept them.
But as soon as we alert them to the fact that there was a lie that was told, people are
very accurate in saying, well, in that case, I think you're the one who actually told this
lie.
Now interestingly, in our initial study, interrogation didn't help.
It was just their thinking back on the whole sequence of events and where the lie likely
was that is what enabled them to discover it.
But nonetheless, it shows that, even when people initially accept something is true,
they're actually quite capable, when the lie is meaningful, the lie is complicated, and
the lie's told to their face, they're capable of going back and finding it.
BLAKE MCKIMMIE: So what did you do in your studies, then, that was different from the
prototypical deception study, that improves deception detection accuracy?
BILL VON HIPPEL: So in this particular study, we did a few things.
First of all, they were always friends with each other.
Because if you're not somebody's friend, you might not know when they're showing cues of
cognitive load or raised pitch of their voice, the things that really are indicative of lying.
Because you don't know what their voice usually sounds like.
And you don't know how quickly they usually speak.
So they are always groups of friends.
And then we had them tell this complicated lie where they're involved with each other,
and they're trying to get their partners to make a bad choice.
It's an ambiguous situation, nobody knows the answer.
But they're pushing an answer that they know is wrong.
But they're told you have to push it subtly.
You'll get paid if your group buys your answer, but only if they don't detect you as a liar.
So doing this as gently as they can, but nonetheless, people know them well, and they get a chance
to think, well, who was speaking more than they ought to if they can go back and look
for unusual events in the episode that took place and try to find the liar or that way.
Real life is often that way.
Of course it's often not.
Maybe the lie took place in hidden circumstances.
But then we can say, well, how well do the other circumstances fit with what the liar's
trying to tell us?
And I think that what these data show us is that interrogation, which we believe is really
effective, may not be terribly effective.
But our good knowledge about people and about the ways they usually behave actually can
be very useful for us detecting, well, when are they trying to do something that's a bit
different from what they would usually do?
BLAKE MCKIMMIE: And so they're able to reasonably, accurately detect deception, even when they're
not actively looking for it just in retrospect, thinking back.
BILL VON HIPPEL: Yeah.
That's exactly right.
So in our case, we had 250 people in this experiment.
And we asked them at the end of this study, once they done their group decision making
task, not knowing that one of the group members was a saboteur, we say to them, OK, what's
this study really about?
It's not about group decision making.
So we were deceiving you.
There's something else going on.
Not a single person out of 250 said it was about deception, and that one of the people
here is a liar.
Then we say, well, it's actually about deception.
One of the people here is a liar.
And when chance would be 50%, if you set it up that way, they were way up, like 75 or
80.
And so the typical the chance is you're a 53.
You're not a 75.
So it's clearly a case where now, even though they believed it the first time, they think
back on all the complicated things that took place during that group discussion.
There's something fishy here.
And Blake, you're the one.
BLAKE MCKIMMIE: Do you think that knowing the person is potentially deceptive is one
of the key factors?
Because you've had feedback from previous interactions with them about when they're
being truthful and not.
Do you think feedback plays a role in increasing their ability to detect deception?
BILL VON HIPPEL: I think that.
But I don't have any evidence for it.
So it's our belief that, in the small groups that we evolved in, we always knew all the
people we interacted with.
So if we were trying to develop abilities, evolve abilities, to detect deception, they're
likely comparing the Blake that I'm watching now with the Blake I've known for a long time.
And making judgments about differences.
And so when somebody is a stranger, you can't do that.
We believe that plays a critical role, but we haven't run the study that involves strangers.
So we don't know.
Even our paradigm wouldn't work with strangers, anyway.
People are awfully polite to strangers, and so they might not be willing to do the things
that they're happy to mess around with their friends.
But it's my suspicion that it matters a lot that you know the person, that we're good
at comparing current to prior behaviors.
That it matters a lot that the lie's important, so there's something really on the line.
That it matters a lot that the lie is complicated.
So there's not just an issue of repeating over and over, "I didn't do it."
BLAKE MCKIMMIE: So to increase the consequences, or to increase the stakes, of getting away
with that deception.
But also in detecting the deception, you paid your participants money.
So can you just tell us a little bit about how that worked exactly
BILL VON HIPPEL: Sure.
So one of the people's the saboteur.
And they're told, we'll pay you for every wrong answer you can convince your group to
use.
The group is paid for every right answer that they choose.
So they're clearly at cross-purposes for each other.
Now, they're not paid much.
A dollar per answer.
The key with the saboteur, though, is that, to really tell a lie, I have to not only convince
you in the moment.
But I have to convince you later on when somebody-- when it discovers that somebody did have an
affair with your wife.
I don't want you to think it's me.
And so they were told, you'll be paid a dollar for every wrong answer you convince your group,
if / when they then find out there's a saboteur, and they don't use you.
And so they could make an extra $10 or $15 if they could successfully lie to their group
and convince the group, even later on when interrogation took place, they're the only
ones who knew, eventually it will be revealed, that there was a liar in here.
And they have to convince the group later on that that liar was not them.
BLAKE MCKIMMIE: Don't you think the money was the only thing that they were really striving
for and trying to get away with the lie?
BILL VON HIPPEL: Look, the money was nice, of course.
But I actually think that what it really came down to was the social fun.
They want to pull one over on their friends and prove that they could lie to them.
And the friends don't want to have somebody pulling the wool over their eyes.
And so the lab is right next to my office, and when it was revealed that there was a
saboteur, there was laughter and accusations and yelling and back and forth.
And it was a lot of fun for them.
And you could see they were bound and determined not to be found out.
And they were also bound and determined to find out who it really was.
And the few dollars probably made little difference.
BLAKE MCKIMMIE: So maybe the conclusion, one thing you might take from your study, then,
detecting deception is really a social thing.
It's not an individual's ability per se.
But it's something that comes out of being in a social group and living in a social context.
BILL VON HIPPEL: That's exactly right.
The deception is a social process, just like truth-telling is.
And so truth-telling is about making sure that you and I both understand the world as
it really is.
And deception is making sure that you and I understand a world that's favorable to me,
but is not how the world really is.
Both of those are social processes.
Sometimes I can get you on board, and you would rather just go with me than really know
the truth.
Because your relationship is more important than the facts.
Other times, the facts are more important than the relationship.
But in all cases, what we're trying to do is create these competing social realities
that allow us to, in the end, understand what's really going on.
But also, in the end, be on the same page as each other.
That's a fundamental human motive
BLAKE MCKIMMIE: Excellent.
Thanks, Bill.
Không có nhận xét nào:
Đăng nhận xét