Charting Quality

An important component of this process is charting the information collected. OmniLingua professionals routinely use a number of different charts to graphically display and analyze statistical information in our efforts to enhance quality. Among those charts and tools used most frequently are:

1. Brainstorming Tools
2. Pareto Charts
3. Cause and Effect Diagrams
4. Run Charts
5. Histograms
6. Control Charts

Brainstorming Tools

Many of the techniques used in Statistical Process Control are designed to precisely focus on a specific aspect of a problem, or what really could be more positively referred to as an opportunity. In essence, isn’t every problem an opportunity to improve?

Brainstorming, while obviously different in its focus than many charting techniques, is an invaluable method of broadening our perspective to view as many aspects of a problem or solution as possible. Whether we use a “structured” approach that systematically requires every member of a group to contribute thoughts to the process; or an “unstructured” approach that is more relaxed and free-flow, and accepts ideas as they surface, there are several guidelines to follow when using this tool:

1. No criticism allowed, there are no bad ideas in a brainstorming session
2. Every idea goes on the board or flipchart, both as a clarification tool, and a memory aid
3. There must be total agreement on precisely what is being “brainstormed”
4. Record what is said, not interpretations
5. Move along smartly, hold as many brainstorming sessions as necessary, but limit them to a quarter hour or less.

Pareto Charts

When a number of problems or conditions are being analyzed, the vertical bar graph format of a Pareto Chart displays all of the items thus providing a clearer perspective of their relative magnitude or importance. Working from this clearer perspective, it is easier to determine which problem to tackle first, or to gain a better feel for root causes of a problem.

Pareto Charts are an excellent tool for focusing our attention on the most important problems – or opportunities.

2Cause and Effect Diagrams

If a Pareto Chart looks at the totality of problems or conditions, a Cause and Effect Diagram (often referred to as a Fishbone Diagram) looks at the various possible causes of those problems or conditions. These diagrams graphically capture the “causes” affecting a process on the left side and state the problem or effect on the right side. Once the various possible causes have been identified, their relative probability can be determined and further study undertaken.

Run Charts

When the desired graphical display of data is to show trends over a finite period of time relative to an “average,” the Run Chart offers an easy-to-create, easy-to-follow solution. The Run Chart provides a visualization of information to help determine if the average is changing. Of significant importance, however, is not to be sidetracked by every variation. Of greater importance is evaluating the trend and statistical validity of events. This should lead to investigation and appropriate action.

Histograms

Just as a Pareto Chart graphically displays the frequency of problems or opportunities in a bar graph, the Histogram is a bar graph showing the distribution of data for a single measured element such as typographical errors. Under normal, repeated conditions a Histogram will have a particular appearance based on the type of data. It is important to analyze each successive Histogram for unexpected results, in particular distribution that varies from the expected pattern.

Control Charts

The control chart is a Run Chart with statistically determined “upper” and “lower” control limits on either side of the “average” line. The Control Chart graphically displays both the variability caused by the system and that caused by special events occurring outside the system. This information helps determine if a process is in statistical control. However, only detailed analysis will reveal whether a process deemed to be in statistical control is acceptable and producing desired results, or merely within control parameters and showing consistency.