Cause-Effect Diagramming: Understanding the Causal Structure in Your Systems

Cause-effect diagrams illustrate the relationships between causes, and between the causes and the net effect. Understanding this ‘causal structure’ is a necessary condition for improving process performance. Without this knowledge process improvement, efforts will wander blindly in search of better performance.

What is a Cause-Effect Diagram?

Ishikawa was a leading and early advocate of, and pioneer of, quality in Japan. He invented the diagram that often bears his name as a simple tool for doing ’cause’ analysis, the search for – and elimination of – the one fundamental cause of quality problems.

Figure 1 shows an example. The ‘Effect’ (Poor Operating Room Turnaround Time) is shown in the box at the right. Major categories of ‘Causes’ (People, Machines, Materials, Measurement, and Methods) form the major ‘bones’ of the diagram, branching above and below the spine. The bones directly connected to these bones are the first level causes, often symptoms of the main problem.

Further branching gets into deeper and deeper levels of cause and illuminates more and more of the causal structure. It is these finer branches that are likely to be a ’cause’ of poor O.R. turnaround time.

Figure 1 - The Traditional Cause-Effect DiagramUsing the Cause-Effect Diagram

One thing the C-E diagram does not indicate is the magnitude of the individual causes. Is, for example, ‘Surgeons Late’ a more important cause than ‘Instruments Not Ready’? The C-E diagram may make a good map, but it lacks time and distance data. For that further data collection is necessary. Checkboxes, histograms, Pareto analysis, and other data collection and analysis tools are used to quantify the occurrence of the individual problems. We will discuss these tools in future editions of Tool of the Week.

Creating the Cause-Effect Diagram

The tool is a simple one and creating one is also simple. First, the major cause categories are identified (the five in Figure 1 are the usual categories, but any number and kind of category are allowed) and within each group, the major causes are delineated. From there the causes of the causes of the causes are discussed and identified.

Another option is to simply brainstorm all of the possible causes and then using a grouping method, such as the KJ method, to assemble the information into the diagram format.

Either of these methods work. The latter method is usually faster; the former method is more detailed.

Enhancing the Cause-Effect Diagram

Process Flow Cause-Effect Diagram

A C-E diagram for an entire process is much too general. To get at specific problems the quality improvement process must focus on all steps in the process individually (or at least the major ones). The Process Flow Cause-Effect Diagram uses the Process Flowchart and adds short C-E diagrams attached to each step in the process. Figure 2 shows a schematic Process Flow Cause-Effect Diagram

This provides the most detailed mapping of the process and its potential problems.

Figure 2 - Process Flow Cause Effect diagram

Shortcomings

One shortcoming of the tool is that it was invented by and for people in process and assembly oriented industries. These processes tend to have simpler and more linear causal structures, a -> b -> c ……. -> z. But many processes are not so simple, a -> b -> c -> d -> a, for example, in a repetitive feedback loop. Or a causes b and c, but c is in a different category than b; where does a go? Or a sometimes causes b and sometimes causes c.

Other cause-effect relationships are even more complex with variations and hybrids of these problems.

Other Cause-Effect Tools

To overcome this difficulty other cause-effect tools are useful. We will save our discussions of the individual tools for later editions, but a very brief introduction will help put them in context.

Peter Senge’s book The Fifth Discipline discusses systems thinking and the diagrams that facilitate systems thinking. Systems thinking provides an excellent tool for exploring and talking about complex systems with nonlinear or dynamic causal structures. One basic cause-effect relationship in systems thinking is the feedback loop. These loops, which can be positive (self-reinforcing) or negative (self-limiting). Deming’s (or Shewhart’s take your pick) famed Plan-Do-Check-Act Cycle at the heart of continuous improvement is an example of a positive feedback loop:

Figure 3 - PDCA as a Positive Reinforcing Causal Loop

In this cycle, a process plan is established, or an experimental hypothesis established, in the Plan stage. The process is implemented during the Do phase and monitored during the Check phase. Based on the results of the checking a decision is made and some process change made, the Act phase. These changes are then incorporated into the new process design during the next Plan stage and the cycle repeats itself, hopefully in a positive cycle of continuous improvement.

Feedback loops can be combined and further influenced by external variables. Senge’s book presents some of these combinations as templates, or ‘archetypes’, from which to build Cause-Effect diagrams of complex systems.

Another tool for cause-effect analysis is Failure Mode and Effect Analysis (FMEA). FMEA, as the name implies, analyzes the effect (or consequences) of problems as well as the relationship between causes and those problems. This, in a sense, is a quantification enhancement of the fishbone diagram. The base tool of FMEA is not the fishbone but, rather, the Tree Diagram. The Tree Diagram contains the same information as the fishbone but in a different format.

A FMEA analysis involves computing a ‘criticality’ index for each cause. This index is based on the estimated likelihood of the problem occurring, it’s severity, and the likelihood of the system detecting (and hopefully preventing) the problem from occurring. This provides a way of quantifying and prioritizing the causes in the diagram.

Summary

While the traditional cause-effect diagram of Ishikawa is an effective tool for many, if not most situations, other tools are useful in analyzing the causal structure of more complex and dynamic systems. What is important is that the tool works and the that the organization gains the necessary process knowledge.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.