Popular in Social Sciences

Books

City Time: On Being Sentenced to Rikers Island

Independent Scholar David A. Campbell

Culture and Society

The City as a Growth Machine: Toward a Political Economy of Place

New York University Harvey Moltoch

Economics

Gender and Economics

University of California Santa Barbara Shelly Lundberg

American Politics

Keeping the Republic

Boston College Marc Landy

Banks and Policy Organisations

Greening the financial sector

European Investment Bank Matteo Ferrazzi

British Politics

Entrepreneurial thinking: the politics and practice of policy impact

University of Sheffield Matthew Flinders

Culture and Society

Carceral Entrapments: Views from the Prison/Street Interface in India

Indian Institute of Technology Delhi Mahuya Bandyopadhya

Planned vs. Actual Attention

View more insights in Social Sciences - Economics - Game Theory

View more figures from Indiana University

View the page of Ala Avoyan

Read the Study

· Tools Cite Embed

Attention allocation in strategic tasks: planned vs actual. Ala Avoyan explores the mismatch and its determinants.

Image courtesy of interviewee. January 8, 2024

Log-in or Sign-up to Faculti
Currently viewing this subject insight as a guest. You have insight(s) remaining for this month.

Workspace

<< >>

Video Summary
Similar Studies
AI Chat
Key Points
Key Quotes
Summary Translate
Full Transcript

Export as PDF

This video video discusses the importance of time allocation in decision-making, emphasizing that time is a scarce and valuable resource. The author and their co-authors conducted studies to investigate how well people can intuitively allocate their time across different tasks or games. They designed an experiment where participants had to plan their time budget between paired games based on their anticipation of time needed. The study highlighted a significant discrepancy between participants' planned and actual attention allocation. It revealed that people struggle to accurately predict the time required for tasks, often focusing too much on potential payoffs during planning rather than the strategic complexities during execution. This video video leads to errors in time allocation, underscoring that understanding where time is most valuable is crucial for effective decision-making. The research suggests that people should identify tasks where planning errors could have the most significant impact and focus their efforts there. Future papers will expand on this by examining individual characteristics like risk aversion and how they influence time allocation strategies.

Language:

Export as PDF

X-TIMESTAMP-MAP:LOCAL=00:00:00.000,MPEGTS=900000

00:00.225 --> 00:01.875
Time is a scarce resource.

00:02.775 --> 00:05.795
If anything in our lives is scarce, it's time.

00:06.615 --> 00:09.115
And because time is so valuable, we would like

00:09.115 --> 00:10.635
to allocate it optimally.

00:11.335 --> 00:15.275
So if we think of our everyday lives, right, we have to make

00:15.335 --> 00:16.395
so many decisions,

00:17.295 --> 00:19.515
and we have to decide how best

00:19.615 --> 00:22.075
to divide our limited time resource

00:22.075 --> 00:23.475
between these different tasks.

00:24.695 --> 00:28.305
This time allocation problem we face can be further

00:28.455 --> 00:30.945
complicated by the fact that the outcomes

00:31.835 --> 00:33.705
don't just depend on our own actions,

00:34.205 --> 00:36.345
but also depend on the actions of others

00:36.845 --> 00:39.425
and maybe even chance, right?

00:41.165 --> 00:46.105
So, you know, when people face all these different problems

00:46.285 --> 00:49.985
of time allocation, how do they go about solving them?

00:50.535 --> 00:52.825
What features of the problems affect

00:52.885 --> 00:54.625
and don't affect people's choices?

00:55.935 --> 00:59.715
Are people good at anticipating the relative amount

00:59.715 --> 01:01.475
of time on a given task?

01:02.735 --> 01:05.075
So my co-authors and I set out

01:05.075 --> 01:08.155
to answer such questions in a se series of studies,

01:08.835 --> 01:12.635
particularly in plan versus actual attention paper.

01:12.855 --> 01:15.275
We investigate how well people can

01:15.745 --> 01:17.635
intuitively allocate their time.

01:18.255 --> 01:21.995
So are people basically good at anticipating

01:22.615 --> 01:26.155
how much time they'll need for one task versus another one?

01:27.105 --> 01:29.885
So this whole thing can be quite complicated

01:29.945 --> 01:31.325
and context dependent.

01:31.825 --> 01:33.845
So the way we go about it is

01:33.845 --> 01:36.245
that we study a stylized version of it.

01:37.165 --> 01:38.265
So we design

01:38.365 --> 01:40.225
and conduct an experiment

01:40.835 --> 01:44.065
where in an incentive incentive compatible manner,

01:44.485 --> 01:49.065
we elicit people's subjective planned attention allocation,

01:50.265 --> 01:54.565
and then using eye tracking data later on, we get their

01:55.645 --> 01:57.815
objective attention allocation,

01:57.815 --> 01:59.455
which we call the actual one.

01:59.475 --> 02:02.015
So we have the planned attention allocation,

02:02.075 --> 02:04.175
and then the actual attention allocation,

02:04.315 --> 02:05.535
we can compare the two.

02:06.805 --> 02:09.985
So, let me give you a little bit more on the

02:10.635 --> 02:12.185
experiment actually.

02:12.405 --> 02:16.945
So what we do, so we present subjects with pairs

02:16.945 --> 02:18.065
of two by two games.

02:18.615 --> 02:19.745
What does that mean?

02:20.285 --> 02:23.625
Uh, we describe games in a form of a matrix, a table.

02:24.085 --> 02:25.665
So there's a game,

02:26.325 --> 02:29.065
and our subjects play with another, partner

02:29.805 --> 02:31.305
who also has two actions.

02:31.335 --> 02:32.465
They have two actions.

02:32.525 --> 02:35.025
So this two by two table tells them, okay,

02:35.365 --> 02:36.945
if I choose option A

02:37.245 --> 02:41.065
and the other person chooses option B, what do I earn?

02:41.205 --> 02:42.905
And what does the other person earn?

02:43.485 --> 02:47.105
We chose to use this visual representation of games

02:47.805 --> 02:51.505
so people have this, easy way

02:51.685 --> 02:54.745
of taking in both games at the same time

02:55.285 --> 02:56.945
and not having any difficulties

02:57.005 --> 02:58.905
of comparing apples and oranges.

02:59.825 --> 03:02.965
So our subjects see a pair of two

03:02.985 --> 03:06.245
by two games on the screen, and they're given 10 seconds.

03:07.345 --> 03:10.365
So in the 10 seconds, we're not telling them

03:10.665 --> 03:11.725
to solve these games.

03:12.275 --> 03:14.525
What we're telling them is, this is a part one

03:14.745 --> 03:15.805
of our experiment.

03:15.825 --> 03:18.845
In part two, you'll have some time budget,

03:19.265 --> 03:22.365
and you'll be playing both of these games right now.

03:22.675 --> 03:26.085
Tell us how would you like to split your total time budget

03:26.155 --> 03:27.365
between these two games?

03:28.305 --> 03:32.445
So they can tell us 19%, 10%, or 50 50

03:33.025 --> 03:35.525
or 25%, 75%.

03:36.485 --> 03:40.325
So 10 seconds is enough time for them to observe the games,

03:40.545 --> 03:42.765
but it's not enough time for them to solve it.

03:43.455 --> 03:46.555
So they're all they're telling us is how they would like

03:46.555 --> 03:49.915
to plan their budget for later on for part two.

03:51.495 --> 03:55.635
And so they do 12 pairs of such games in the part one,

03:56.265 --> 03:57.555
they just plan their time.

03:57.585 --> 04:00.755
They look at them for 10 seconds, they give us a split.

04:01.865 --> 04:03.035
Part two comes around.

04:03.735 --> 04:05.675
One of the 12 pairs is realized,

04:06.415 --> 04:08.475
and they have to play both of the games

04:08.855 --> 04:11.955
and the time they have follows exactly the

04:11.955 --> 04:13.115
time that they told us.

04:13.895 --> 04:17.755
So if they split the, their time, 50 50 half

04:17.755 --> 04:20.555
of the time is given for the first game

04:20.615 --> 04:22.675
and the rest goes for the second game.

04:23.695 --> 04:27.115
So now comes the part three.

04:27.895 --> 04:30.735
So now we tell them part one

04:30.755 --> 04:33.375
and part two of the experiment is over.

04:34.425 --> 04:39.405
You can move on and earn a little more money by doing what?

04:39.535 --> 04:41.245
There are 11 pairs of games

04:42.455 --> 04:44.275
for which we have their plan time,

04:44.735 --> 04:46.755
and now we tell them there are no time limits.

04:47.385 --> 04:50.915
Just go ahead and solve those games

04:51.095 --> 04:52.995
and for we will pay you for one of them.

04:53.975 --> 04:56.755
So what's happening in the background is

04:56.755 --> 04:58.435
that we're tracking their eyes

04:59.045 --> 05:00.995
throughout this entire experiment.

05:01.455 --> 05:02.755
And this is quite neat.

05:02.755 --> 05:05.355
There's like this little device that sits on the,

05:05.455 --> 05:07.395
the monitor and tracks their eyes.

05:07.505 --> 05:11.555
It's really an intrusive, and so they go ahead

05:11.575 --> 05:13.475
and do the 11 pairs of games.

05:14.335 --> 05:16.235
And during that we can calculate

05:16.295 --> 05:19.995
how much time they actually spend on each game by seeing

05:20.065 --> 05:21.555
what they see on their screens.

05:21.695 --> 05:23.515
So we we'll calculate that,

05:23.775 --> 05:26.955
and now we have their actual time for 11 pairs.

05:27.615 --> 05:29.195
We also have their plan time.

05:30.055 --> 05:33.515
So the, the key here is then comparing their plan time

05:33.775 --> 05:34.955
to the actual time.

05:35.295 --> 05:37.195
And then we can see the, the,

05:37.195 --> 05:39.475
if there are any discrepancies, basically.

05:41.405 --> 05:45.205
So the what we find is quite interesting.

05:45.345 --> 05:49.405
So what we find is there is a drastic difference between

05:49.435 --> 05:51.165
what people think they're gonna do

05:51.705 --> 05:53.205
and what they actually do.

05:54.385 --> 05:55.845
Uh, we find that plan

05:55.865 --> 05:58.845
and actual attention allocation differ, differ,

05:59.225 --> 06:03.045
and people are not particularly skilled at instinctively

06:03.135 --> 06:06.445
allocating their time across games in a manner

06:06.915 --> 06:08.605
that correctly anticipates

06:08.665 --> 06:11.365
how much time they'll ultimately take to decide.

06:12.175 --> 06:14.835
And the rest of the paper is devoted to

06:15.515 --> 06:17.835
identifying why there is such a mismatch.

06:18.105 --> 06:20.435
Because without it, we just said, okay,

06:20.585 --> 06:23.555
there's one more thing that people might be not great at.

06:24.575 --> 06:28.315
So, to we di dive more into this,

06:28.455 --> 06:31.075
and we draw from one of our previous papers.

06:31.415 --> 06:35.955
So Andrew Shorter and I have a, a paper from 2020.

06:36.615 --> 06:39.915
Um, attention in Games was published,

06:39.975 --> 06:43.435
in European Economic Review in which,

06:43.975 --> 06:46.475
we do this part one, part two

06:46.575 --> 06:48.675
of our experiment did I described today.

06:48.775 --> 06:50.515
So that's all about planning,

06:51.255 --> 06:55.115
and we were interested in what determines,

06:55.315 --> 06:56.635
people's planning choices.

06:57.305 --> 06:59.955
What features of games makes people think

06:59.955 --> 07:02.755
that one game is more valuable than the other one.

07:03.255 --> 07:06.035
And there we identify that the,

07:06.185 --> 07:09.835
what people focus most on is the best case scenario

07:10.695 --> 07:12.195
and the worst case scenario.

07:12.855 --> 07:14.115
So people look at the games

07:14.495 --> 07:15.715
and they ask themselves,

07:15.855 --> 07:18.955
so what's the best thing I can get in this one game?

07:20.205 --> 07:22.505
If it's a larger number, they think

07:22.505 --> 07:23.865
that game is more valuable.

07:24.765 --> 07:27.535
This a similar thing goes for the worst case scenario.

07:27.535 --> 07:30.455
They want the worst case scenario to be better.

07:31.475 --> 07:34.495
So if the worst case scenario is not so bad,

07:35.045 --> 07:38.135
that game gets more time allocated to it.

07:38.905 --> 07:43.125
So people avoid things like, zero payoffs

07:43.125 --> 07:44.765
and they shy away from those.

07:45.345 --> 07:46.805
So that's while planning.

07:48.105 --> 07:51.285
So in the current paper, what we identified is very similar.

07:51.705 --> 07:53.685
We replicated our previous results

07:53.865 --> 07:57.805
and when planning people really focus on the, the payoffs,

07:58.815 --> 08:03.225
but when they play the games, they're more concerned

08:03.225 --> 08:04.945
with strategic complexity.

08:06.615 --> 08:10.555
So the differences that we see between the planned

08:10.615 --> 08:12.395
and actual attention arise

08:12.395 --> 08:15.955
because the modes that people use, the heuristics

08:15.955 --> 08:17.475
that they use when they plan

08:17.615 --> 08:19.915
and they when they play, are very different.

08:20.345 --> 08:22.955
When they plan, they mostly focus on payoffs.

08:23.015 --> 08:24.355
What's the best thing they can get?

08:24.355 --> 08:25.915
What's the worst thing they can get?

08:26.295 --> 08:30.115
Are there some zeros to avoid when they play?

08:31.365 --> 08:33.375
It's more about strategic complexity.

08:33.515 --> 08:36.695
And strategic complexity really,

08:38.485 --> 08:40.525
describes well the time that they spend on it.

08:40.865 --> 08:42.645
So we have two different modes,

08:43.455 --> 08:46.075
and of course those different modes lead

08:46.075 --> 08:48.435
to different results that then don't match.

08:49.095 --> 08:52.235
So those are kind of, the main results

08:52.265 --> 08:54.715
that we get and it's nice to

08:55.985 --> 08:57.805
Put our finger on it and say why.

08:58.385 --> 09:03.205
So people overestimate the value of games

09:03.585 --> 09:05.125
by looking at the best

09:05.145 --> 09:07.645
and worst things that they can get when they plan.

09:08.305 --> 09:11.605
But those they, things might not be relevant when they play

09:11.605 --> 09:14.845
because they could be dominated by other actions.

09:14.945 --> 09:18.485
So the strategic complexity really comes in,

09:18.585 --> 09:21.085
and the actual attention we've established

09:21.085 --> 09:22.445
that people's plan

09:22.585 --> 09:26.125
and actual, allocation could be quite different,

09:26.665 --> 09:30.045
but you might be asking yourself, why should we care

09:30.105 --> 09:32.245
and when do we should care about it?

09:33.065 --> 09:37.365
Uh, and to really answer this question, well,

09:37.385 --> 09:40.245
we have this experiment too in our paper

09:40.775 --> 09:44.885
where we allow our subjects to play, the same game

09:45.425 --> 09:47.205
for a long period of time, and we,

09:47.505 --> 09:51.085
and they're allowed to change their mind as they go along.

09:51.625 --> 09:55.925
So what this allows us to do is that we can see how

09:56.445 --> 09:58.885
frequently they change their A answers

09:59.185 --> 10:00.325
and if they change at all.

10:01.105 --> 10:02.605
So what this tells us is

10:02.605 --> 10:05.845
that some games are more time sensitive than others.

10:06.315 --> 10:08.515
What does that mean? You might

10:09.345 --> 10:11.035
make some bad planning choices,

10:11.895 --> 10:13.835
but they could be inconsequential.

10:14.495 --> 10:16.555
But then there's some other games where

10:17.315 --> 10:19.595
planning mistakes could be drastic.

10:20.775 --> 10:24.705
So what we're learning from our paper is that we need to

10:25.265 --> 10:29.425
identify where your time is the most valuable.

10:30.625 --> 10:33.935
Where is it that the planning errors could make the

10:33.935 --> 10:34.975
most impact?

10:35.555 --> 10:39.575
And we need to allocate our time in, in those tasks.

10:41.155 --> 10:44.975
Uh, so we have follow up papers that will be coming again

10:46.335 --> 10:47.485
later this year

10:47.935 --> 10:51.765
where we focus more on people's characteristics such

10:51.765 --> 10:56.565
as risk aversion and ambiguity, attitudes and confidence

10:56.665 --> 11:00.725
and say, well, given such set of characteristics,

11:00.995 --> 11:04.085
what should be people focusing on and whatnot.

11:05.025 --> 11:08.805
Uh, but in this paper, the the big conclusion is that

11:10.445 --> 11:15.215
find the tasks where your time makes the most difference

11:15.875 --> 11:16.895
and focus on those.

Faculti AI can make mistakes