Scriptapedia/Building a CLD with Paper

From Wikibooks, open books for an open world
< Scriptapedia
Jump to: navigation, search


Building a Causal Loop Diagram with Paper

This script is used to capture the variables and causal structures that emerge during a participant discussion early in the GMB process before any structures or diagrams have been made. This script can overlap with the "Graphs Over Time" script or any discussion during which participants are discussing relationships within the system.

Status:

Promising practices

Primary nature of group task:

Convergent

Time:

Preparation time: 10 minutes

Time required during session: 40 minutes

Follow-up time: 15 minutes

Materials

  1. Pad of paper or multiple sheets of paper for each modeler
  2. Pencils or pens
  3. Large whiteboard, chalkboard, or flipchart paper to present initial diagram

Inputs:

Open discussion during which participants discuss variables within the system and links between variables (e.g., during the "Graphs Over Time" script)

Outputs:

Causal loop diagram

Roles

  • Recorder who is able to draw causal linkages as they emerge during the discussion and is able to identify variables
  • Modeler with expertise in system dynamics who is able to quickly stitch together causal linkages to form a feedback model, and who also has experience in editing diagrams in system dynamics to make them correct
  • Facilitator to guide discussion among participants

Steps

  1. Participants are engaged in a discussion, guided by the facilitator, about variables and behaviors within the system of interest. This type of storytelling could take place through a number of scripts, including the "Graphs Over Time" script.
  2. As the discussion proceeds, the modeler writes down variables named by participants and the causal links between variables on sheets of paper.
    • Each link should relate at least two variables with a directed arc and identify the link as either positive or negative.
    • Delays should also be indicated if they are mentioned.
  3. Meanwhile, the recorder identifies variables named by participants and writes one variable in marker on each index card/post-it note.
  4. If there are additional recorders, they can assist the modeler by also documenting causal loops on paper.
  5. The modeler begins to create a complete causal loop diagram (CLD) based on the linkages identified on the sheets of paper. The modeler will often redraw the entire structure to produce a better layout.
    • Be careful not to add links or variables that are not explicitly stated by participants.
  6. Take a group break.
    • During a break, the modeler and recorders discuss the variables and causal loops they have identified, making sure there is consensus on language used and connections made.
    • As a team, they identify 5-6 key variables with causal linkages. These index cards/post-it notes are then placed/taped on chalkboard or whiteboard. Causal arrows with polarities that emerged from the participants' discussion during Part A are drawn between the variables.
    • All the other index cards/post-it notes with the other variables are placed in columns beside the initial causal loop diagram.
  7. After the break, the modeler reviews the diagram with all of the participants, starting with, “This is what we heard you saying…” From there, the modeler walks through the small CLD.
  8. The modeler takes care to explain that:
    • A ‘+’ sign means same effect, so that an increase in one variable leads to an increase in the other, and a decrease in one variable leads to a decrease in the other.
    • A ‘-‘ sign means an opposite effect, so that that an increase in one variable leads to a decrease in the other variable, and a decrease in one variable leads to an increase in the other.
    • The ‘+’ and ‘–‘ signs are not “good” or “bad.” They just reflect the direction of change.
    • The concept of reinforcing and balancing feedback loops should also be introduced and illustrated with the CLD diagram.
  9. As the modeler introduces major SD terms (e.g. feedback loop, +, - ), the terms should be written up on a flipchart/board by the facilitator along with a brief definition so that all participants can see and refer to it.
  10. The facilitator then tells participants, “This is only the beginning of a much larger system. During the conversation before the break, we were listening for key variables and we’ve written them all down on index cards. Now we want you to continue building this map of relationships. For example, how does X variable [choose an index card from the columns] fit into this picture?” The facilitator then invites participants to come up and place the variable on the board, and draw an arrow connecting it to the existing CLD.
    • Whenever a participant identifies an additional variable or a connection between variables, the facilitator invites him/her to stand up, add the index card to the CLD and physically draw the modifications on the model.
  11. The facilitator and modeler alternate inviting participants up to the board and coaching on polarities.
    • If time is running out or saturation has been reached, the facilitator asks for one or two last suggestions, and then stops the exercise.
  12. The modeler will close with a summary of major feedback loops and key variables, and will highlight model-based insights that emerge.
  13. The facilitator then guides groups in reflecting on the process:
    • What has this conversation and CLD brought up?
    • What are the implications for how you’re going to approach the problem?
    • What have you learned about the system?
    • How do you want to move forward?
  14. The recorder photographs the CLD diagram or collects the flip charts. The modeler transfers the CLD to Vensim.

Evaluation Criteria:

  • Participants created a rich causal loop diagram based on their thoughts and stories with the aid of the modeling team
  • CLD diagram created and drawn in Vensim

Authors

Annaliese Calhoun and Timothy Hower, 2010

History

Based on Building a CLD from Discussion

Revisions

None

References

None

Notes

None