# Discrete Vs. Volume

Six Sigma – iSixSigma Forums Old Forums General Discrete Vs. Volume

Viewing 7 posts - 1 through 7 (of 7 total)
• Author
Posts
• #49233

Swaggerty
Participant

If I have a large volume of count data, do I consider that to be continous. e.g. No. of defective parts / Total parts manufactured. In both the cases (numerator & denominator) my data is discrete. If I were to look at the % defective parts, will be data be considered continuous?
Will the volume of the underlying data determine my data type? If volume is > 1000 then % measure is continuous and if < 1000 then % measure is discrete?

0
#167992

Taylor
Participant

George,
Your Data is Descrete, this you know. Once you convert the data to a percentage such as good/bad  then you are talking about variable Binary Descrete data, other wise having a value rank or order. If no value rank or order can be assigned, then the data is considered Attribute Binary Descrete.(Example Male/Female) Continuous data cannot exist in this form It is only continuous when reference to a scale, length, or time etc. (Ratios and Intervals)
There is no underlying volume that determines Descrete/Continuous data.
Hope this helps

0
#167993

Swaggerty
Participant

What if I measure total number of defective parts every day as a % measure. E.g. % defective parts manufactured everyday for 3 months. I will be measuring % over a large variable volume on a continual basis. In this case will it be considered continuous / discrete?

0
#167994

Taylor
Participant

George
Your data is still discrete, However if you report your data as a function of time lets say %defects per month then one might consider it to be Continuous, but its still Discrete Variable Binary Data.
Continuous Data Examples are something like this

How long it takes to take a test
Temperature of the sun
Weight of a Child when born
Voltage output of a car battery

0
#168022

6Sigma Below The Belt
Participant

Your data is a rate and can be considered pseudo-continuous provided you have the following:
1. the denominator is >100 or better if >1000
2. Only 20% of the total data gathered should have the same value
3. p or np >5 or 95

0
#168125

DD
Participant

If 1000 pepole lie instead of 100 that does not make it the truth.

0
#168134

Sayyed
Participant

Thanks 6Sigma Below the Belt:
I now recall, these guidelines are similar to what my mentors had instructed during one of my GB projects. So, just to confirm – when my data is pseudo-continuos I will be able to use all the tests/tools that assume Y data to be continuous, right?
Thanks again, – Neal

0
Viewing 7 posts - 1 through 7 (of 7 total)

The forum ‘General’ is closed to new topics and replies.