Motivation Barriers - Question for Nick

[Jochen Fromm-3]

Someone asked recently in my discussion group if you can use
multi-agent based modeling to examine the behavior of humans.
His name is Pieter Steenekamp (pieter_steenekamp at yahoo.com)
and he is working for a health care service provider. The
objective of his work is to influence the patients to change their
lifestyles in order to live a healthier life; to motivate them
in the most cost effective way to exercise more, stop smoking,
loose weight, follow a healthy diet etc.

People usually need a motivation and an incentive for something,
they choose the most pleasant and comfortable things, and this
is of course not always the healthiest choice. If they have
the choice, most people will choose normally the Whopper with
Cheese and Bacon and a large strawberry milk-shake instead of
the healthier rucola salad with crispbread and mineral water.
Once they get used to a healthier lifestyle, they might enjoy it,
though.

Nick, you are a Psychology professor and certainly interested
in multi-agent based modeling of human behavior ? Do you
think a few simulations about this subject would be interesting ?
An example for such a simulation is for instance a system
with selfish agents which select normally the comfortable,
pleasant actions which "feel good". These agents must learn
to select the incomfortable, unpleasant actions which "feel bad"
in the short term but are good for the agents in the long term.
A way to do this are for example catalysts: the group in form of
peer pressure ("weight watchers") or money in form of incentives.

It is well known that the offering of incentives can cause also
all forms and types of unintended consequences. People will
do everything to get the incentives: if a government for
instance offers money for snakes in order to reduce a snake
plague, people may start to breed snakes instead. Would it
be interesting to simulate the overcoming of motivation
barriers with agent-based modeling ?

-J.

[Stephen Guerin]

Hello Jochen and Peter,

Nick doesn't stay up late -- I thought I would sneak in a response before he
wakes up on the East Coast. :-)

>From the description you forwarded, I don't immediately see the need to model a
large population of agents and their interactions to get at the questions being
asked. Agent-based modeling might not be the best fit.

The system so far described might best be modeled with a systems dynamics
approach (a system of coupled differential equations). Different qualities of
food could be modeled as input flows and different health measures could be
modeled as stocks which are a partial function of these inputs. The probability
of healthy food consumption could be a function of incentives and peer pressure.
So far, not a very complicated model and probably with no great surprises...
Adding feedback loops or time-lagged variables are two ways to get more
complicated behavior out of this type of model.

One place I could see ABM contributing would be looking at the opinion dynamics
in the peer pressure you mentioned or how health fads might spread through a
social network. Mike's work with drug epidemics comes to mind.

Looking for unintended consequences is interesting. Though writing an open-ended
model that can generate new solutions is difficult to do without cheating or
'pre-baking' in the solutions in the possible model outcomes. Genetic algorithms
and genetic programming offer some possible pathways for generating novelty in
models but it would require a large corpus of realworld facts to generate even
the trivial example of people deciding to breed snakes when there's an incentive
to kill them.

That's my .02 anyway...Nick, are you awake yet?

-Steve

> -----Original Message-----
> From: Jochen Fromm [mailto:fromm at vs.uni-kassel.de]
> Sent: Tuesday, February 07, 2006 2:09 AM
> To: 'The Friday Morning Applied Complexity Coffee Group'
> Subject: [FRIAM] Motivation Barriers - Question for Nick
>
>
>
> Someone asked recently in my discussion group if you can use
> multi-agent based modeling to examine the behavior of humans.
> His name is Pieter Steenekamp (pieter_steenekamp at yahoo.com)
> and he is working for a health care service provider. The
> objective of his work is to influence the patients to change their
> lifestyles in order to live a healthier life; to motivate them
> in the most cost effective way to exercise more, stop smoking,
> loose weight, follow a healthy diet etc.
>
> People usually need a motivation and an incentive for something,
> they choose the most pleasant and comfortable things, and this
> is of course not always the healthiest choice. If they have
> the choice, most people will choose normally the Whopper with
> Cheese and Bacon and a large strawberry milk-shake instead of
> the healthier rucola salad with crispbread and mineral water.
> Once they get used to a healthier lifestyle, they might enjoy it,
> though.
>
> Nick, you are a Psychology professor and certainly interested
> in multi-agent based modeling of human behavior ? Do you
> think a few simulations about this subject would be interesting ?
> An example for such a simulation is for instance a system
> with selfish agents which select normally the comfortable,
> pleasant actions which "feel good". These agents must learn
> to select the incomfortable, unpleasant actions which "feel bad"
> in the short term but are good for the agents in the long term.
> A way to do this are for example catalysts: the group in form of
> peer pressure ("weight watchers") or money in form of incentives.
>
> It is well known that the offering of incentives can cause also
> all forms and types of unintended consequences. People will
> do everything to get the incentives: if a government for
> instance offers money for snakes in order to reduce a snake
> plague, people may start to breed snakes instead. Would it
> be interesting to simulate the overcoming of motivation
> barriers with agent-based modeling ?
>
> -J.
>
>
> ============================================================
> FRIAM Applied Complexity Group listserv
> Meets Fridays 9a-11:30 at various locations
> lectures, archives, unsubscribe, maps at http://www.friam.org
>
>

[Michael Agar]

Gruessdich Jochen. Your colleague might enjoy taking a look at the  
Plexus Institute web page, www.plexusinstitute.org, if he doesn't  
know about it already. It's a center for links between health care  
and complexity in general.

One day I'd like to work on the obvious relation among models of  
diffusion of innovation, epidemiology, memetic theories of culture,  
consumer behavior etc etc and apply them to social change,  
particularly "health promotion," as your colleague's interests would  
be called here. Whether an agent based model would play a role once  
the theoretical synthesis was in place would, as always, depend on  
whether there was a clear and simple enough argument to test.  
Complexity seems to base itself on the idea that simple rules produce  
complex patterns. ABM's seem to be the way to test whether, given  
interest in a pattern, you can identify the simple rules that produce  
it.

So I'd look at some historical cases of dramatic social change in  
health behavior and see if I could find simple rule changes that  
explained the S curve. When I think of one interesting case in the  
US, the decline in smoking cigarettes (but not among all groups), it  
looks like analysis of such cases would be a major project.

Psychology alone wouldn't do it, though. (Boy, is Nick going to have  
a weird morning when he reads these messages). Most health promotion  
I know of still relies on a rational actor model, just put the true  
propositions in there and the actor will do what's right. We need  
other levels, downscale to explain, say, addiction, and uphill to  
explain things like social conventions, policy, and the markets, and  
then sideways to include other aspects of the person.

Besides, a bacon cheeseburger versus rucola and crispbread? (:

Mike

On Feb 7, 2006, at 2:08 AM, Jochen Fromm wrote:

>
> Someone asked recently in my discussion group if you can use
> multi-agent based modeling to examine the behavior of humans.
> His name is Pieter Steenekamp (pieter_steenekamp at yahoo.com)
> and he is working for a health care service provider. The
> objective of his work is to influence the patients to change their
> lifestyles in order to live a healthier life; to motivate them
> in the most cost effective way to exercise more, stop smoking,
> loose weight, follow a healthy diet etc.
>
> People usually need a motivation and an incentive for something,
> they choose the most pleasant and comfortable things, and this
> is of course not always the healthiest choice. If they have
> the choice, most people will choose normally the Whopper with
> Cheese and Bacon and a large strawberry milk-shake instead of
> the healthier rucola salad with crispbread and mineral water.
> Once they get used to a healthier lifestyle, they might enjoy it,
> though.
>
> Nick, you are a Psychology professor and certainly interested
> in multi-agent based modeling of human behavior ? Do you
> think a few simulations about this subject would be interesting ?
> An example for such a simulation is for instance a system
> with selfish agents which select normally the comfortable,
> pleasant actions which "feel good". These agents must learn
> to select the incomfortable, unpleasant actions which "feel bad"
> in the short term but are good for the agents in the long term.
> A way to do this are for example catalysts: the group in form of
> peer pressure ("weight watchers") or money in form of incentives.
>
> It is well known that the offering of incentives can cause also
> all forms and types of unintended consequences. People will
> do everything to get the incentives: if a government for
> instance offers money for snakes in order to reduce a snake
> plague, people may start to breed snakes instead. Would it
> be interesting to simulate the overcoming of motivation
> barriers with agent-based modeling ?
>
> -J.
>
>
> ============================================================
> FRIAM Applied Complexity Group listserv
> Meets Fridays 9a-11:30 at various locations
> lectures, archives, unsubscribe, maps at http://www.friam.org

[Jochen Fromm-3]

 
I have thought for example of the sugarscape model
from Joshua Epstein and Robert Axtel. You could
imagine for example some agents on a landscape
with red dots (the whoppers with high immediate
reward but low long-term benefits) and green
dots (the crispbread with low immediate reward
but high long-term benefits). The preferences and
habits could be changed by a kind of reinforcement
learning, and the interaction could for example include
trade and gossip (for the spreading of fads). How
do you motivate the agents to consume the unpleasant
green dots with long-term benefits instead of the
pleasant red dots with short-term benefits ?

One could for instance introduce a third kind of
intermediate resource, yellow dots. If the
agents get used to this kind of dots, they might
easier switch to the healthier green dots, because
they "look" similar, and they are used to consume
the yellow dots. Another possibilites is to simulate
some form of peer pressure, perhaps by introduction
of some reputation of image score (red dots = high
immediate reward but low image score, green dots =
low immediate reward but high image score).

The success of a method could be measured by the
average life expectancy of the agents, which could
in turn depend on their lifestyle (consumption of
many red dots = lower life-expectancy, consumption
of many green dots = higher life-expectancy).

-J.

[Robert Holmes-2]

I remember seeing this last year in The Times:

http://www.timesonline.co.uk/article/0,,8127-1084068,00.html

All about how the Finnish government got a public health program to work and
transformed a nation of potential heart-attack victims to a bunch of buff
nordic-skiing types. Article makes some mention of incentives (including
cash!)

Robert

On 2/7/06, Jochen Fromm <fromm at vs.uni-kassel.de> wrote:

>
>
> Someone asked recently in my discussion group if you can use
> multi-agent based modeling to examine the behavior of humans.
> His name is Pieter Steenekamp (pieter_steenekamp at yahoo.com)
> and he is working for a health care service provider. The
> objective of his work is to influence the patients to change their
> lifestyles in order to live a healthier life; to motivate them
> in the most cost effective way to exercise more, stop smoking,
> loose weight, follow a healthy diet etc.
>
> People usually need a motivation and an incentive for something,
> they choose the most pleasant and comfortable things, and this
> is of course not always the healthiest choice. If they have
> the choice, most people will choose normally the Whopper with
> Cheese and Bacon and a large strawberry milk-shake instead of
> the healthier rucola salad with crispbread and mineral water.
> Once they get used to a healthier lifestyle, they might enjoy it,
> though.
>
> Nick, you are a Psychology professor and certainly interested
> in multi-agent based modeling of human behavior ? Do you
> think a few simulations about this subject would be interesting ?
> An example for such a simulation is for instance a system
> with selfish agents which select normally the comfortable,
> pleasant actions which "feel good". These agents must learn
> to select the incomfortable, unpleasant actions which "feel bad"
> in the short term but are good for the agents in the long term.
> A way to do this are for example catalysts: the group in form of
> peer pressure ("weight watchers") or money in form of incentives.
>
> It is well known that the offering of incentives can cause also
> all forms and types of unintended consequences. People will
> do everything to get the incentives: if a government for
> instance offers money for snakes in order to reduce a snake
> plague, people may start to breed snakes instead. Would it
> be interesting to simulate the overcoming of motivation
> barriers with agent-based modeling ?
>
> -J.
>
>
> ============================================================
> FRIAM Applied Complexity Group listserv
> Meets Fridays 9a-11:30 at various locations
> lectures, archives, unsubscribe, maps at http://www.friam.org
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://redfish.com/pipermail/friam_redfish.com/attachments/20060207/39d03e3d/attachment.htm

[Douglas Roberts-2]

It's not stated explicitly in the article, but java is not on the list of
languages used...

--Doug

On 2/7/06, Robert Holmes <rholmes62 at gmail.com> wrote:

>
> I remember seeing this last year in The Times:
>
> http://www.timesonline.co.uk/article/0,,8127-1084068,00.html
>
> All about how the Finnish government got a public health program to work
> and transformed a nation of potential heart-attack victims to a bunch of
> buff nordic-skiing types. Article makes some mention of incentives
> (including cash!)
>
> Robert
>
> On 2/7/06, Jochen Fromm <fromm at vs.uni-kassel.de> wrote:
> >
> >
> > Someone asked recently in my discussion group if you can use
> > multi-agent based modeling to examine the behavior of humans.
> > His name is Pieter Steenekamp (pieter_steenekamp at yahoo.com )
> > and he is working for a health care service provider. The
> > objective of his work is to influence the patients to change their
> > lifestyles in order to live a healthier life; to motivate them
> > in the most cost effective way to exercise more, stop smoking,
> > loose weight, follow a healthy diet etc.
> >
> > People usually need a motivation and an incentive for something,
> > they choose the most pleasant and comfortable things, and this
> > is of course not always the healthiest choice. If they have
> > the choice, most people will choose normally the Whopper with
> > Cheese and Bacon and a large strawberry milk-shake instead of
> > the healthier rucola salad with crispbread and mineral water.
> > Once they get used to a healthier lifestyle, they might enjoy it,
> > though.
> >
> > Nick, you are a Psychology professor and certainly interested
> > in multi-agent based modeling of human behavior ? Do you
> > think a few simulations about this subject would be interesting ?
> > An example for such a simulation is for instance a system
> > with selfish agents which select normally the comfortable,
> > pleasant actions which "feel good". These agents must learn
> > to select the incomfortable, unpleasant actions which "feel bad"
> > in the short term but are good for the agents in the long term.
> > A way to do this are for example catalysts: the group in form of
> > peer pressure ("weight watchers") or money in form of incentives.
> >
> > It is well known that the offering of incentives can cause also
> > all forms and types of unintended consequences. People will
> > do everything to get the incentives: if a government for
> > instance offers money for snakes in order to reduce a snake
> > plague, people may start to breed snakes instead. Would it
> > be interesting to simulate the overcoming of motivation
> > barriers with agent-based modeling ?
> >
> > -J.
> >
> >
> > ============================================================
> > FRIAM Applied Complexity Group listserv
> > Meets Fridays 9a-11:30 at various locations
> > lectures, archives, unsubscribe, maps at http://www.friam.org
> >
>
>
> ============================================================
> FRIAM Applied Complexity Group listserv
> Meets Fridays 9a-11:30 at various locations
> lectures, archives, unsubscribe, maps at http://www.friam.org
>
>

--
Doug Roberts, RTI
505-455-7333 - Office
505-670-8195 - Cell
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://redfish.com/pipermail/friam_redfish.com/attachments/20060207/008605f0/attachment.htm

[Jochen Fromm-3]

It seems to be an actual problem.
The Guardian has this week a story about "health labels":
http://www.guardian.co.uk/food/Story/0,,1705639,00.html

The United States face the same problem
http://news.bbc.co.uk/2/hi/americas/4501646.stm 

I think color-coded traffic-light labels would work best.
Detailed informations about calories and how much fat,
sugar and salt a product contains is not read by anyone.

-J
________________________________

From: Robert Holmes
Sent: Tuesday, February 07, 2006 2:43 PM
To: The Friday Morning Applied Complexity Coffee Group
Subject: Re: [FRIAM] Motivation Barriers - Question for Nick

I remember seeing this last year in The Times:
http://www.timesonline.co.uk/article/0,,8127-1084068,00.html

All about how the Finnish government got a public health program to work and
transformed a nation of potential heart-attack victims to a bunch of buff
nordic-skiing types. Article makes some mention of incentives (including
cash!)