Scatter Diagram...

Information | Understanding | Best Practice.


Scatter diagrams can be utilized to analyze two variables to determine if there is any form of relationship between them. If a relationship is identified, then the possibility arises that one variable may be controlled by varying the other variable. The strength of the scatter diagram lies in simplicity, however, it is important that all other potential variables within a process are understood, controlled and monitored to ensure the results obtained from any experiments or interpretation of a diagram are not compromised.

Scatter diagrams are also known as Scatter Plots or X-Y Diagrams.


Scatter diagram example.

In the following example we are investigating and seeking to control the output from a process where there is a high speed drilling machine. We want to determine if the speed of the drill (drill speed) used in the process, impacts the smoothness of the final cut finish on a metal bar. If it turns out that the smoothness of the final cut is related to the drill speed, the desired finish can be controlled by means of controlling the drill speed. For this example we might construct a graph of drill speed versus smoothness of cut. For each variable there will be an expected range, i.e. drill speeds may range from “low” to “high” and “finish of cut” may span a range from “rough” to “very smooth”.

Scatter Plots

Quality Management Tools and Techniques …
Information | Understanding | Best Practice   >>>

In preparing to create the scatter diagram, we will identify all potential variables associated with the process. This may take the form of performing a “Brainstorm Analysis”, where individuals such as machine operators, support engineers, maintenance technicians, line supervisors, internal customers of the process, all suggest possible sources of variation within the drilling process. The 6M’s normally applied in performing a “Cause and Effect”‘ can form a solid basis for the Brainstorm session. Once all the variables have been identified, it will be important to either control these variables or monitor them during the scatter analysis.

Following on from the review of the potential sources of variation, the next step will be to agree a test plan. An approach may be to perform a range of tests over a period of time, for example over different work shifts. It is important to ensure the materials used during the analysis are consistent and stable. The planned approach should be documented and agreed up front by all involved.

With the plan agreed, then proceed to perform the testing.  Run batches of product through the drilling process and record the smoothness achieved versus the drill speed. The information gathered can now be plotted on a diagram and we may see a trend as follows.

Quality Management Tools and Techniques …
Information | Understanding | Best Practice   >>>

In this example graph, we see as the drill speed is increased, the smoothness of finish also improves. This result would be very easy to interpret and the appropriate instructions could easily be implemented via documenting into a “Standard Operating Procedure”.
A more realistic result may be in the form of the following diagram, where we see the smoothness improve up to a point, then as we continue to further increase the drill speed, the smoothness of cut achieved starts to dis-improve.

Scatter Plots

Quality Management Tools and Techniques …
Information | Understanding | Best Practice   >>>

Where there is a relationship identified between the two variables a “best fit line” may be applied to the graph, in order to permit the development of an equation to represent the variables relationship.


Where can scatter diagrams be usefully applied?

Scatter diagram analysis can be utilized:

As part of a “cause and effect” analysis. A scatter diagram analysis can be applied to verify a cause and effect relationship.
The scatter diagram can simply and clearly demonstrate the effect, of any process change or process improvement activity.
When applying process control charting, the scatter diagram can indicate the presence of an autocorrelation.
The relationship between two variables can help point to “root cause” of problems, during a root cause analysis.
By performing multiple scatter diagram analysis, the existence of numerous effects from a single source cause can be identified.


How to create a Scatter Diagram.

a) Identify the two variables which you want to investigate.
b) Setup a data collection process. It is important to ensure that all other potential variables remain constant. If these variables cannot be controlled, then they need to be monitored.
c) Start to record data for both set of variables. The greater the number of data points collected, the better the resulting diagram will be.
d) If the variables normally operate over a range, try to collect data over as much of the range as possible.
e) As you collect data, start to plot on a scatter diagram.
f) Review the diagram. Determine if a relationship exists.


It is common practice to consider the two variables to act along the lines of being an independent variable and a dependent variable. The independent variable is normally plotted along the X-axis (horizontal), and the dependent variable is plotted along the Y axis (vertical). The thinking being that as the independent variable changes, the dependent variable moves in accordance with how it displays on the scatter diagram.

Caution needs to be applied when building a scatter diagram, in that “correlation does not imply causality”, i.e. a relationship may be identified between two variables, however it must not be assumed that one variable is causing an effect with the other variable. There may be other variables inputting into the process and impacting the results, therefore, it is important to control and monitor all other potential variables, to investigate and fully understand potential cause and effect relationships identified via the scatter diagram analysis.


TQM Tools and Techniques Full Details

Quality Improvement Techniques

Total Quality Management Tools and Techniques …

        • Continuous improvement utilizing Analytical Techniques.
        • Brainstorming
        • 5 why’s analysis
        • Process Flow Diagrams/Flowcharts/Process Mapping
        • Check sheets /Check Lists
        • Run charts
        • Histograms
        • Scatter Diagrams/Scatter Plot
        • Cause and Effect/Fishbone/Ishikawa Diagrams
        • Identifying sources & causes of variation
        • Control/Shewart Charts/DPU Charts
        • Cpk and Ppk Analysis
        • Pareto Analysis
        • Bottleneck Analysis
        • Benchmarking
        • FMEA
        • FTA
        • HAZOP
        • SIPOC
        • Etc. Etc.
        • Information | Understanding | Best Practice   >>>