Six Sigma is data-driven. Six Sigma practitioners who have completed the Lean Six Sigma Green Belt training or other Lean Six Sigma Green Belt courses will be aware that Six Sigma teams face many types of data in different units. Relative measures are measures of the variance in a range of values, regardless of their unit of measurement. This allows you to directly compare the spread of two values with different measures with relative dispersion. This information is particularly useful during the Measure and Analyze phases in the DMAIC process.
Participate in our 100% online and self-paced Six Sigma training.
Types of Relative Measures for Dispersion
There are four relative measures for dispersion.
Coefficient of Quartile Deviation
Coefficient of Mean Deviation
All the relative measures of dispersion are known as coefficients. We will only be discussing three of the four: coefficients of range and quartile deviation.
Relative Measures Of Dispersion: The Coefficient of Range
This relative measure of dispersion is based on the range value. This is an example of a relative measure of dispersion, also known as the ‘Range Coefficient of Dispersion’. It is the largest value minus the smallest value divided with the largest value plus the smallest value. Let’s take a look at this figure for an illustration.
Let’s take two sets. Set A contains marks from seven Geometry students out of 25 marks, while group B contains marks from the same number of Mathematics students out of 100 marks. The range of marks in each subject will be calculated. The absolute range in Geometry is 11. While the absolute range in Mathematics is 26. This is based upon absolute measures of dispersion and not relative measures of it. However, the truth is that they cannot be directly compared as their bases are different. These two values can be converted into coefficients of range. The coefficient of range for Geometry is higher than the one for Mathematics. Geometry has more variation and dispersion. Mathematics students have more stability than Geometry students. This is learned using relative measures of dispersion.
Relative Measures Of Dispersion: Coefficient Quartile Deviation
How do you calculate the coefficient of deviation quartile?
Let’s take the example of geometry and mathematics marks. We will use relative measures to disperse the data concerning quartiles. Now, we will calculate the coefficients of quartile departure for mathematics and geometry using Q3 minus Q1 divided with Q3 plus Q1. We see that the coefficients of quartile departure for Geometry and Mathematics are similar. It is 0.5 for both subjects. The inference is that both subjects have the same median performance.
None of the subjects showed a higher or lower uniformity of median scores than any other. Remember the fundamental rule regarding relative measures of dispersion. A small coefficient of quartile deviation indicates high uniformity. You can find the fundamental rule about relative dispersion here. A small coefficient of quartile deviation indicates high uniformity, a small variation in the central 50% of items, or high uniformity towards median performance. If the coefficient of quartile variance is high, it indicates that there is a lot of variation among the central 50 items or that the median performance is uniformly lower. This was the second relative measure of dispersion.
Relative Measures Of Dispersion: The Coefficient of Variation
Let’s take a look at the last relative me.