Histograms
Histograms provide a graphic summary of variations in a set of data. This is done by partitioning the range of data into several intervals of equal length, counting the number of points in each interval, and plotting the counts as bar lengths. Histograms are useful in the study of process capability since they graphically display one or more of three important data distribution properties: shape, location, and scatter.
Typically, the shape of the distribution should be normal or bell shaped. Any significant deviation from the normal pattern has a cause which, once determined, can shed light on the variability in the process.
Histogram analysis is a basic step in analyzing a process and can provide the following sense of accomplishment:
- We have quantified some aspect of the process; we are managing by facts, not opinions.
- We have a better understanding of the variability inherent in the process; we have a more realistic view of the ability of the process to produce acceptable results consistently.
- We have new ideas and theories about how the process operates or about the causes of a problem and we have set the stage for additional investigative efforts.
See Also
Creating and Viewing Histograms
