In an earlier post “Data Collection? That’s One Tough Job, Buster” (http://qualityg.blogspot.com/2005/07/quality-tooltechnique-data-collection.html) I stated I would write more on Data Collection.
The following is an accumulation of information (different sources) I learned over the years, the application of this information and the methods used are mine. I hope it can help others.
One of the most overlooked criteria’s required prior to data collection activities is to create an
“Operational Definition.”
An Operational Definition is a precise description that tells how to get a value for the characteristic you are trying to measure. It includes what something is and how it is to be measured. The key word is characteristic (s). It is also important to understand that you may need a separate definition for each subject area. For example, Employee “A” may have a different definition than Employee “B”. The purpose of an Operational Definition is two-fold:
- To remove uncertainty so everyone has the same understanding
- To make sure that no matter who does the measuring, the results are always the same The following is an example Operational Definition I used when I worked in the Voucher Department:
“An error-free expense report includes receipts for all items on the Expense Form over $25.00 and does not include any receipts for items not included on the Expense Form. There are no math errors and expenses are coded properly.”
- To remove uncertainty so everyone has the same understanding
- To make sure that no matter who does the measuring, the results are always the same The following is an example Operational Definition I used when I worked in the Voucher Department:
“An error-free expense report includes receipts for all items on the Expense Form over $25.00 and does not include any receipts for items not included on the Expense Form. There are no math errors and expenses are coded properly.”
or,
“The starting time to measure customer response time is when the customer enters our store and when the customer leaves.”
Think of the Operational Definition in Data Collection as you would the Problem Statement for Root Cause Analysis. Both should be stratified (i.e., broken/categorized) to provide a focus that can be measured and understood by all involved. Without this we will probably conduct meaningless data collection studies and reports (unfortunately I’ve done that many times when pressured for data).
Do NOT allow this process to be come bureaucratic, a nice guide (checklist) to follow that I inherently use when doing data collection or when I have no idea where to start aquality conversation is I ask myself Who, What, Where, When, Why and How (by “What Method”):
1. Clarify Data Collection goals
• Decide why you are collecting data
• Decide what data you need to collect
Decide how much (sample size) data you need to collect (i.e., are you looking for point in time or for an on-going trend)
• Decide What you will do with the data once you have it
2. Develop an Operational Definition
• Define what you are trying to evaluate
• Decide how you will attach a value to what you are trying to measure (develop data collection form and corresponding data base if necessary)
• Decide how you will display/record the data (i.e.; graph/chart, spreadsheet, report, etc.)
• Determine the period of time you will conduct the data collection activities
3. Test for Data Consistency
• Determine factors that may cause your data to vary from one item to another (i.e., people, your own subjectivity, location, time, size, different process, etc.)
• Reduce impact of those factors
4. Data Collection
• Train all involved in study to ensure consistency (use associated data collection forms and input data base if required)
• Make data collection procedures error-proof
• Have frequent process checks (do a sample test on small group) to determine if data collection forms and tools are accurate and appropriate (i.e.; does the data look reasonable and actionable)
By following these guidelines you will not only improve your data collection activities, you will also reduce cycle time in analyzing and displaying your data with the appropriate graphs and charts.
p/s -
When you can measure what you are speaking about and express it in numbers, facts, theory or application, then you know something about it, but when you cannot measure it, envision it, or provide meaningful information about it, or if you cannot express it in numbers, your knowledge or so called expertise is of an insufficient and unacceptable type.
No comments:
Post a Comment