Anticlue

Measuring with Continuous Data or Discrete Data

As you go about transforming data into information one begins to realize the importance of how the data was measured. The Six Sigma measurement technique refers to the type of data chosen to be measured within a process.

The data that you measure for a Six Sigma project is especially important because it can indicate the size and seriousness of the problem in the process you are investigating, which will help you improve that process. Different measurements are taken to gather different data. The length of time for a process to execute, the count of defects produced in a day, or the dimension on a product part are all different types of measures. When one decides to measure, understanding the type of data to be measured is important for assuring accuracy of the measurement. The two categories of data are:

• Continuous data – Continuous data is data that can assume a range of numerical responses on a continuous scale. Continuous data is also known as variables data. Since continuous data can measure a wide range of values, the slightest defect can be measured and detected.

  The two types of continuous data are:

  1. Physical property data - Physical property data depicts the tangible aspects of a process. Examples of physical property data are weight, width, length, height, and temperature. All of these examples can be divided into smaller measures. A meter can be divided into centimeters.
  2. Resource data - Resource data details the assets related to a process, such as time and money. Both Time and money can also be sub-divided into ever-smaller segments.
  
  Physical property data and resource data are types of continuous data because they can be measured on an infinitely divisible scale. This ability can be used to illustrate the smallest levels of deviation.

  As with everything, continuous data has its pros and cons. On the positive side, continuous data may inform one the degree to which the product has missed the mark, and the small sample sizes may be adequate for evaluating the impact of a change in a key process input variables. Hover on the negative side, continuous data is complex and requires a measuring instrument which adds to the cost of gathering the data.

• Discrete data – Discrete data is the presence or absence of some characteristic in each device under test. Discrete data is also know as attribute data and it has an either/or quality.

  The three types of discrete data are:

  1. Characteristic data - Characteristic data details the attributes of a process. An example of characteristic data is measuring how many networking break/fixes rather than desktop break/fixes are being resolved each day.
  2. Count data - Count Data depicts the number or frequency of an observable event that occurs in a process. Commonly used to measure defects.
  3. Intangible data - Intangible data illustrates a piece of the process that is intangible. For example, a person's feeling might be measured on a scale of good to bad, or high to low. Intangible data is being measured in determining, on a scale of 1 to 5, business customers' satisfaction with their IT department. The artificial measurement of 1 to 5 is necessary because there is no other way to measure this intangible trait.
  
  Discrete data also has its positive and negative points. For the advantages, discrete data is quick and easy to collect due to its either/or characteristic. Discrete data is easily understandable. Everyone gets a customer satisfaction rating when they are given the scale. However, discrete data may require a large sample size to identify changes in process input variables, and the nuances of the data tend to get lost in the size.

Both categories of data, continuous and discrete, have advantages and disadvantages. For Six Sigma continuous data is normally preferred. Determining what data is needed is depended on the process being audited.