Step 3A: Choosing the Correct Control Chart (Discrete Data)

If the data type that needs to be charted is discrete, then it must fall between one of binary or count types. As the name suggests in case of binary distribution, there are only two possibilities, success and failure, defective and not defective, whereas in the case of count type distribution there may be more than two possibilities of the process, the number of defects still need to be counted. The difference between the two is subtle. However, given the fact that there are fixed rules governing when which chart needs to be used, the confusion is reduced.

Count

Equal Subgroup Size: In this case, the Six Sigma process methodology suggests using the C Chart. The C Chart counts the number of defects that are happening per unit of unit. This could be per minute, per hour, per day or per week and so on. The important part here is that since the time period is fixed, so is the sample size. The C Chart will tell you how many observations from a given sample failed to meet the criteria specified by the control limits.

Unequal Subgroup Size: In this case, the Six Sigma process methodology recommends using the “U” chart. The U chart counts the rate of defects. This means that it keeps a track of the number of units that it has monitored and how many have failed the given criteria specified by the control limits. The U Chart will tell you, for example that 5/1253 units have failed to comply. Hence it does not require a fixed time period or a fixed sample size and may be more convenient to use.

Binary

Equal Subgroup Size: In this case, the Six Sigma process methodology recommends using the NP Chart. The NP chart also counts the number of defects per period of time just like the C Chart. There is however a subtle difference between the two. C Chart is used when the occurrence of defects is rare. However, NP chart uses the binomial distribution. Therefore, the occurrences do not have to be rare. As a thumb rule, if there is a chance more than 5% defects in a process, then the NP chart must be used.

Unequal Subgroup Size: In this case, the P chart is recommended. Once again, it is quite similar to the U chart. This chart calculates the defects as a percentage figure meaning that it takes into account the units that have gone through the process just like the U Chart. The U chart also uses the Binomial distribution and is used when the occurrences of defects is not rare.

Similar Articles Under - Six Sigma - Control Phase

View All Articles