|
Collecting Quality Data
The Three Kinds of Data
To improve quality, we must collect and analyze a broad range of information or data in order to understand and fix an operating process - its inputs, throughputs and outputs. Because data itself varies from instance to instance, we call it a variable (as opposed to a constant, which is unchanging, and therefore not worth tracking because its presence is taken for granted). There are three different kinds of variables that we may have to track to improve quality. A particular improvement effort may require that we collect and examine data of all three kinds, or perhaps only one kind. These are;
numerical variables, expressed as numbers,
categorical variables, where values fall within predetermined categories, &
rank variables, which convey relative degrees within a certain range
A particular variable may simultaneously belong to two or even all three of these kinds or categories, depending on how you look at it. For example, a test score of 75% may be conveyed as the numerical variable 75, or a categorical variable if 75% is deemed to be a B+, or even a rank variable where the number is used to determine the relative position of that value in comparison to others. The kind of variable we are tracking will determine how we track it, and how we analyze it.
Picking the Right Tool
The five data collection tools described below allow us to gather quality data of all kinds. But they are not only for inspectors or others who are charged with ensuring quality. Quality must become a routine part of everyone's job, and therefore everyone must be able to use these tools in their work. Indeed, the true value of these tools becomes apparent only when front-line operators know how to use them, and put them to work to monitor and fix or balance a process. These tools provide operators with the insights they need to improve quality, and a language they can use to communicate effectively with others about their efforts.
We use Dot Plots to track numerical or rank data which fall within a smaller range of possibilities.
When we have to track numerical or rank data that fall within a larger range of possibilities, we use the Stem and Leaf Chart.
To track the incidence of categorical data, we use a Tally Sheet.
To track the incidence of categorical data, visually, we use Check Sheets.
To track the trends of data over time, we use Time Plots.
|
|
