Cautions on using existing data

Using existing data lets you take advantage of archived data or current measures to learn about the output, process or input. Collecting new data means recording new observations (it may involve looking at an existing metric but with new operational definitions).

Using existing data is quicker and cheaper than gathering new data, but there are some strong cautions:

The data must be in a form you can use
Either the data must be relatively recent or you must be able to show that conditions have not changed significantly since they were collected
You should know when and how the data were collected (and that it was done in a way consistent with the questions you want to answer)
You should be confident that the data were collected using procedures consistent with your operational definition
They must be truly representative of the process, group , measurement system
There must be sufficient data to make your conclusions valid

If any of these conditions are not met, you should strongly think about collecting new data.

Tip

It's seldom wise to use old data only. Existing data is best used to establish historical patterns and to supplement new data.

Making a checksheet

Highlights

Design a new checksheet every time you collect data (tailored to that situation)
Having standard forms makes it easy to collect reliable, useful data
- Enables faster capture and compiling of data
- Ensures consistent data from different people
- Captures essential descriptors (stratification factors) that otherwise might be overlooked or forgotten

To create and use a checksheet …

Select specific data and factors to be included
Determine time period to be covered by the form
- Day, week, shift, quarter, etc.
Construct the form
- Review different formats on the following pages and pick one that best fits your needs
- Include a space for identifying the data collector by name or initials
- Include reason/comment columns
- Use full dates (month, date, year)
- Use explanatory title
- Decide how precise the measurement must be (seconds vs. minutes vs. hours; microns vs. millimeters) and indicate it on the form
  - Rule of thumb: smaller increments give better precision, but don't go beyond what is reasonable for the item being measured (Ex: don't measure in second a cycle time that last weeks—stick to hours)
Pilot test the form design and make changes as needed
- If the "Other" column gets too many entries, you may be missing out on important categories of information. Examine entries classified as "Other" to see if there are new categories you could add to the checksheet.
- Make changes before you begin the actual data collection trial

Basic checksheets

	Week
Defect	1	2	3	4	Total
Incorrect SSN					3
Incorrect Address					1
Incorrect Work History					2
Incorrect Salary History					8

Easy to make and use
Simply list the problems you're tracking and leave space to allow marks whenever someone finds that problem
The example shown here also includes a time element

Frequency plot checksheet

Easy to do by hand while a process is operating

Repair shop output rate (Jul 1–Jul 19)
Date	Completed repairs
1
2	X	X	X	X	X	X	X
3	X	X	X	X	X
4	X	X	X	X	X
5	X	X	X	X
6	X	X
7	X	X	X
8	X
9	X	X	X	X	X	X
10	X	X	X	X
11	X	X	X	X
12	X	X	X	X
13	X
14	X	X	X
15
16	X	X	X	X	X	X
17	X	X	X	X	X
18	X	X	X	X	X	X	X	X
19	X	X	X	X

Automatically shows distribution of items or events along a scale or ordered quantity
Helps detect unusual patterns in a population or detect multiple populations
Gives visual picture of average and range without any further analysis

Flylib.com

Category list

Similar pages

Cautions on using existing data