sytems thinking

7/31/2019 sytems thinking

1/16

A Primer on Systems Thinking & Organizational

Learning

John J. Shibley /The Portland LearningOrganizationGroup

Home

Making Loops: A Method for Drawing Causal Loop Diagrams:

IntroductionThis article describes a method for drawing Causal Loop diagrams,of going from a story in which one "smells a loop" to a fully formed

Causal Loop. This is not an article for innocents. It assumes you

have a basic understanding of reinforcing and balancing structures,

and that you've struggled with trying to make Causal Loop

diagrams.

This method grew out of a mutual inquiry with my clients in the

Systems Thinking Learning Group at Columbia St. Mary's

Hospital in Milwaukee. They wanted to know how I "did it" - how Imade loops from the stories they told me.

I realized I was interested in this too, and so we began a process of

watching me work, developing a model about what I was doing,

and refining this model through several iterations of further

observation and model making.

Here's the method...

1. Sense a story

2. Listen to the story

3. Listen to the story again, with "variable ears"

4. Create variables

5. Create links

6. See if the links make a loop
http://www.systemsprimer.com/jsbio.htmhttp://www.systemsprimer.com/index.htmlhttp://www.systemsprimer.com/jsbio.htmhttp://www.systemsprimer.com/index.html


2/16

There are also two general rules that permeate the method:

Rule #1: As you move through the model, freely use what

you learn to revisit and refine work done in earlier stages.Rule #2: During it all, attend to the assumptions being

made, and the way those assumptions are formed out of

data.

No good method is complete without a picture. Here's the picture

we created...

Beginning with The Story, one proceeds through a process of

Identifying the variables involved

Establishing the links between variables

Creating loops based on these linksThis process is shown in the heavy lines since it is the

primary path forward toward a finished causal loop.

The lighter lines illustrate that the process is iterative

(Rule #2), and requires constant attention to the

assumption making process (Rule #1).

The rest of this article explains each step in more detail.

I'd be very pleased if this piece did one of the following for you.

Maybe, just maybe, the way I "do" loops will help

you do loops.


3/16

Even if my approach doesn't work for you, I hope it

is helpful for you to see how one person approaches

this challenge.

Most importantly, I hope the following encourages

you to reflect on how you do whatever you do, tobecome a more reflective practitioner.


About the StoryThe first three steps in this process are all about the story.

Step 1: Sense a story.

The first step in creating a causal loop diagram is

suspecting that there might be one in the first place!Everyday we are barraged by dozens of stories that

could potentially be described using causal loops. The

ability to "sense" when there may be a systemic

structure driving the results is one quality of a

developed systems thinker.

I call this "system sensing" because in my experience

the realization that a loop may be hiding in a story is

not so much a thought ("Oh, I bet there's a loop here.")

as it is an intuition. It's like a gentle tickle somewhere

in your consciousness. This tickle, if attended to, will


4/16

lead to the thought "Oh, I bet there's a loop here."

However, the thought is not first, the tickle is.

This "system sensibility" is probably the hardest task

for the beginning system thinker to understand, letalone develop. I'm not sure how one develops it other

than practicing drawing causal loops over and over. I

think it's a little like learning to be a musician. The

best musicians have a musical "sense" of what will

work, but that sense usually develops only after many

years of heavy lifting like scale work and practice.

In any case, I do believe people can become better at

sensing the system thinking tickle, first by simplyknowing it's there, and secondly by listening for it.

Step 2: Listen to the story

Once you sense that there may be a "loop-able" story

listen to the story. Just listen. Listen without trying to

find the archetype or the loop or any of that. Just

listen.

How come? Two reasons.

1. It's really good discipline. Most of us

want to leap into action, to pursue our

insights the way a dog runs after a bone.

The consequence can be that we end up

hearing and drawing the story we tell

ourselves instead of the story our client

tells us.

2. Drawing loops is a rational process

informed by intuition, and intuition, as

any good psychotherapist will tell you,

requires the involvement of the

unconscious. The unconscious is a


5/16

necessary partner to the rational, but

will often not participate unless invited.

By hearing the story once with a blank

mind, we invite our unconscious to

become involved.

OK, so you've heard the story. Now what?

Step 3: Listen to the story again, with "variable ears".

The second step is "Listen to the story again, with

'variable ears'". Have the story teller tell the story

again, and this time listen with your attention

particularly attuned to things in the story that are

increasing or decreasing, trending in one direction orthe other or perhaps both at different times.What's a variable?

Variables are those things in the story that change,

that increase or decrease over time. A sound loop is

built on sound variables. If the final loop you create

will be the graphic equivalent of the story you've

heard, the variables are the carefully chosen words of

that story.

For example, consider this story...

"This stock market is really something,

eh? Stock prices are high, so people feel

really optimistic, and so they buy more,

which drives stock prices even higher.

That makes people even more

optimistic, so they buy even more stock,driving prices higher again."

In this story, there are three things that are changing

over time, all increasing: stock prices, market

optimism and buying. Those are the variables in this

story.


6/16

If you let the story teller know that you're paying

attention to the variables in the story, you may find

that they tell the story with a "variable tongue"

(yuck!). They will begin to use language like "this

goes up" or "this goes down" as they retell the story.They're own understanding of the story will begin to

shape itself into the raw material of a causal loop.

Notice at this point that you're still ONLY listening. In

fact, this may be the third time you've heard the story.

That's OK. I find it sometimes takes three iterations

before I've quieted the noise in my own head enough

to really hear the story in my client's head.


VariablesStep 4: Create variables.

Once you've heard the story again, move to

identifying the variables that might be in play. The

drive to make the variables right inevitably causes us

to return to the story itself, and deepen our

understanding of it. That's why the picture above hasa light path leading from the Variables back to The

Story(Not sure what a variable is? Go back one page to that

section.)
http://www.systemsprimer.com/making_loops_storysteps.htm#What%E2%80%99s%20a%20variable?http://www.systemsprimer.com/making_loops_storysteps.htm#What%E2%80%99s%20a%20variable?http://www.systemsprimer.com/making_loops_storysteps.htm#What%E2%80%99s%20a%20variable?http://www.systemsprimer.com/making_loops_storysteps.htm#What%E2%80%99s%20a%20variable?


7/16

In reality, there are two separate steps here.

The first step is getting potential variables public. I

like to write variables on large post-its and stick them

to a flipchart easel, in what amounts to a brainstormmethod. It makes it very easy to move the variables

around once they're created, and saves a lot of rewrite

time.

For a first pass, I simply record the variables without

worrying about how "good" they are as variables.

Let's look for the variables in our example of the stock

market.

The StoryThe

Variables

"This stock market is really

something, eh? Stock prices are

high, so people feel really

optimistic, and so they buy more,

which drives stock prices even

higher. That makes people evenmore optimistic, so they buy even

more stock, driving prices higher

again."

Stock PricesBuyer

Optimism

Buying

The second step is to get very precise about what

each variable should say, which is really a process of

helping people clarify their thinking.

I find these challenges in creating variables...

Being "tight": A variables should be

measurable. Ideally, you should be able

to create a trend chart of the values

described in the variables.


8/16

Think back to our example of the stock

market. Each of those variables (stock

prices, buyer optimism and buying )

could be measured.

Being "right": The variable needs to

describe the thing that actually changes.

This can prove more challenging than it

sounds. Often you have to work with

people to help them think carefully

about what exactly increases or

decreases.

For example, take the variable "buying"in our stock market story. Does the

storyteller mean the market volume, the

number of investors, or the ration of

buying to selling? Each may be a

legitimate variable, which is true to the

story?

Being prolific: Often at this stage in

developing the causal loop people willsee a great many potential variables,

obviously far more than will be useful in

creating an elegant, descriptive loop. I

usually err on the side of inclusion,

It's also helpful to state the variable in neutral

language. For example, the variable "Stock prices"

could be stated as "Increase in stock prices". However,

the loop should be as easy to understand with themarket increasing or decreasing, and it takes some

mental gymnastics to figure out what a "decrease in

the increase" is (say THAT three times fast!). It's much

simpler to state the variable as "Stock prices" and

have those mistakes decrease or increase.


9/16

With our example of the stock market story, let's

assume that after working the variables we settle on

these three:

For "Stock Prices" I'd probably suggest that the "Dow Average" is

really what is being referred to here, and see ifthe story teller would agree to it. If not, we'd

keep working, because it's their story.

For "Buyer

Optimism"

This one is tricky. There is an Index of

Consumer Confidence that measures this. I'd be

inclined to leave the variable as it is,

acknowledging that we have no exact measure

of it, though we could create one if we had to.

For "Buying" Trickier still. It seems to me that what the story

teller is really talking about here is the

behavioral response to an increase (or decrease)

in optimism. It seems to be about demand for

stocks, since increased demand is what drives

prices upward. (Note that a decrease in demand

for stocks would also decrease stock prices.)

Let's settle on the variable "Demand for Stocks"


Links, S's and O'sStep 5: Create links

Once you've got the key variables described closely

enough, ask people to link the variables described. By


10/16

this time you've probably got an idea about where to

start. Put that post-it on a clean piece of flip chart

paper.

Two valuable questions are "What does this lead to?",which gets the arrow moving forward, and "What

leads to this?" which helps people identify the

preceding variable.

When we add links between variables, we are making

visible our assumption that there is a causal the

relationship between them, that one impacts the next.

This process of clarifying different assumptions is one

of the most useful results of a causal loop exercise.

For example, in this link...

...we're asserting that a relationship

exists between the descriptiveness of

the TV violence rating system and the

absolute amount of violence on TV.

Translated into English, we

would read the above link

like this...

"A change in the

descriptiveness of the TVviolence rating system will

cause a change in the

amount of violence on

TV."

However, different people will assume different

things about this relationship; some claiming that the

amount of violence on TV will increase as a result of

the descriptiveness of the rating system, someclaiming that it will decrease.

In causal loop grammar, we say that one variable

causes the next to change in either the same or the

opposite manner. If the change is in the same


11/16

direction, we label it with an "S". If the change is in

the opposite direction, then we label it with an "O".

If you believe that an increase in the

descriptiveness of the TV violencerating system will lead to a decrease in

the amount of TV violence, then put an

O on the link, like this...

This link reads...

A change in the rating system will

cause an opposite ("O") change " in

TV violence".

This is the argument based on the belief TV shows

will try to avoid more extreme ratings, the way that

movies work to avoid an X rating, because it will

drive away audiences.

The counterarguement is that violence and sex

attracts viewers, which will lead to high ratings,

which will lead to higher levels of violence as other

shows strive to compete.

That link would be labeled like

this...

As a sentence, it would

read...

"A change in

the rating

system will

cause the

same ("S")

type of

change in the

amount of


12/16

violence on

TV."

(Return toDrew's Note)

In our stock market example, we would end up with

these links.

A change in the demand for stock

causes the same (S) kind of change

in the Dow Average. That is, as

demand increases the Dow

Average increases in turn.

A change in the Dow Average


in Buyer Optimism. As the Dow

rises, or falls, Buyer Optimism will

do the same.

A change in Buyer Optimism


in Demand for Stock. As

Optimism falls, so will Demand.

You inevitably make false starts and errors at this

point, pursuing a chain of causal events only to

realize you've headed down a blind alley. This is not

an indication that something is "going wrong". Quite

the contrary : understanding what is not going on is

an important step toward building an accurate

picture of what is happening.


Creating Loops, and So What?Step 6.: See if the links make a loop
http://www.systemsprimer.com/drewsnote.htm#Return%20from%20links%20referencehttp://www.systemsprimer.com/drewsnote.htm#Return%20from%20links%20referencehttp://www.systemsprimer.com/drewsnote.htm#Return%20from%20links%20reference


13/16

This step is not absolutely distinct from the one

above. (As you can see, It grows out of the creation of

a set of links that inevitable feedback on itself to make

a whole. In reality, if the initial steps have been done

well, the creation of a descriptive loop usually

happens naturally and easily.

In our stock market example, we would end up with

this picture.

The Story"A change in the

demand for stock causesthe same kind of change

in the Dow Average.

This in turn causes the

same type of change in

buyer optimism, which

causes the same type of

change in the demand

for stock."

Notice that this loop will "work" in either a rising or

falling market. This is one of the ways that you know

you've drawn the loop in a technically correct way.

So what?


14/16

OK, so you've drawn a loop. Big deal. How does it

help you?Loops help you in two ways.

First, the process of getting to a loop that everybody

agrees on forces people to make explicit their

assumptions. That's the reason for the rule...

Rule #2: During it all, attend to the

assumptions being made, and the way those

assumptions are formed out of data.

One's view of systemic dynamic reflects one's own

mental models. As the process above is unfolding,

one will stumble over these assumptions. The work of

creating variables or links drives people back to

clarify their story, and that work in turn forces people

to make explicit the assumptions their story is based

on, and the data that led them to those assumptions.

One's view

of systemic

dynamicreflects one's

own mental

models. As

the process

above is

unfolding,

one will

stumble over

theseassumptions.

The work of

creating

variables or

links drives


15/16

people back

to clarify

their story

This work, in

turn, forcespeople to

make explicit

the

assumptions

their story is

based on,

and the data

that led them

to theseassumptions.

In our stock market example, it may be that some

people feel strongly that hard economics drive the

market far more than the mood of buyers, while

others may see market behavior as more sociological,

as in the example. The process of articulating the

story, variables and links somehow makes it possible

to expose differences in a rational way, in a way that

leads to wondering how other's see it differently,

rather than attempting to prove them wrong.

Second, the loop helps you understand the way that

the elements in the system are reinforcing or

balancing one another. In the stock market example

we can see that the three variables interact in a way

that keeps them all increasing, or in a bear market,

decreasing.Once we know what is really happening, we can

invent responses that have more chance of

succeeding. Another article on this site, A Practice

Theory for Organizational Learning , describes a

Learning Action Matrixdescribing 4 different "zones"

of work that groups can move through to integrate
http://www.systemsprimer.com/writings.htm#A%20Practice%20Theory%20for%20Organizathttp://www.systemsprimer.com/writings.htm#A%20Practice%20Theory%20for%20Organizathttp://www.systemsprimer.com/matrix_explained_two.htm#The%20Learning%20Action%20Matrixhttp://www.systemsprimer.com/matrix_explained_two.htm#The%20Learning%20Action%20Matrixhttp://www.systemsprimer.com/matrix_explained_two.htm#Four%20Zones%20of%20Workhttp://www.systemsprimer.com/matrix_explained_two.htm#Four%20Zones%20of%20Workhttp://www.systemsprimer.com/writings.htm#A%20Practice%20Theory%20for%20Organizathttp://www.systemsprimer.com/writings.htm#A%20Practice%20Theory%20for%20Organizathttp://www.systemsprimer.com/matrix_explained_two.htm#The%20Learning%20Action%20Matrixhttp://www.systemsprimer.com/matrix_explained_two.htm#Four%20Zones%20of%20Workhttp://www.systemsprimer.com/matrix_explained_two.htm#Four%20Zones%20of%20Work


16/16

action and reflection. Included in that piece are a set

of rules to help people redesign systems.

For some more tips on drawing causal loops, gohere.

For more about feedback, read An Introduction to

Feedback by Leslie Martin at theMIT SystemsDynamics Road Map site.

Acknowledgments

I want to give special acknowledgment to the

Columbia/St. Marys Hospital Systems Thinking

Study Group in Milwaukee, Wisconsin, and to

Stephanie Speer for her passion about story telling.
http://www.systemsprimer.com/detail_three_and_four.htm#Changing%20Systemshttp://www.systemsprimer.com/detail_three_and_four.htm#Changing%20Systemshttp://www.systemsprimer.com/tips.htmhttp://www.systemsprimer.com/tips.htmftp://sysdyn.mit.edu/ftp/sdep/Roadmaps/RM2/D-4691.pdfftp://sysdyn.mit.edu/ftp/sdep/Roadmaps/RM2/D-4691.pdfhttp://sysdyn.mit.edu/road-maps/home.htmlhttp://sysdyn.mit.edu/road-maps/home.htmlhttp://sysdyn.mit.edu/road-maps/home.htmlhttp://www.systemsprimer.com/detail_three_and_four.htm#Changing%20Systemshttp://www.systemsprimer.com/detail_three_and_four.htm#Changing%20Systemshttp://www.systemsprimer.com/tips.htmftp://sysdyn.mit.edu/ftp/sdep/Roadmaps/RM2/D-4691.pdfftp://sysdyn.mit.edu/ftp/sdep/Roadmaps/RM2/D-4691.pdfhttp://sysdyn.mit.edu/road-maps/home.htmlhttp://sysdyn.mit.edu/road-maps/home.html

Documents

sytems thinking