Staying Afloat in a Sea of Data 1

So Much Data!

During the Six Sigma Measure phase of your project you will probably be collecting a great deal of data.

If you aren’t careful, you’ll end up collecting a lot more than you need (either by chasing too much data initially, or by having to go back during the Analyze phase to capture the data you should have gotten the first time).

A quick refresher of the different types of data can help you focus your attention on what is most likely to benefit your process improvement project.

Types of Data:

1. Verbal Data
   1. These often come from surveys: “Nice blog, but I wish this post had been about Data from Star Trek.  Or the Goonies.”
2. Numeric Data
   1. Attribute Data – Discrete variables
      1. Categorical:
         1. Binary: “yes/no”
         2. Nominal: “walk, drive, or fly” (no natural order)
         3. Ordinal: “very satisfied, satisfied, neutral, dissatisfied, very dissatisfied” (natural order)
      2. Count:
         1. Number of occurrences: “number of flaws in a product”
3. Measurement Data – Continuous variables
   1. “infinite” values
   2. Degree of granularity is limited only by the measuring process used
4. Other Data:
   1. Cross-sectional
   2. Time series

The type of data collected influences the statistical analysis tools you can use to analyze that data.  Some data types are more valuable information than others, and some are harder to collect.

It is important to understand what you are really measuring in order to ensure you capture the data that is most appropriate to analyze your process.  Considering that some data is more expensive to collect than others, the cost benefits of different types of data can influence our measurement choices.

No list is complete.  Are these the only data types you use, or have you used others?  Share your stories by leaving a comment below!