Defining what Statistical Tools are - Is a Check Sheet a Statistical Tool?

[sakkapak]

As it mentions in TS 16949 requirement
7.5.1.3 The organization shall use statistical methods of
verification where applicable.

and

8.1.1 Appropriate statistical tools for each process shall be determined during advance quality planning and included in
the control plan.

Could the word "statistical tools and method" be considered as a check sheet?

My opinion is "check sheet" can be considered as a basic statistical tools, since, we utilize the data (mean, standard deviation, ....) for decision making.

Thank for your comment

 

[D.Scott]

IMHO, a check sheet is certainly a quality tool. I wouldn't consider it a "statistical" tool if you were just checking off items that needed to be done. This would indicate more of an attribute was being recorded - "Done or Not Done"

If you were collecting numerical data to be analyzed, I would imagine it is no longer a check sheet. There would be more to it than that. For example, you would be entering the data into some additional system for breakout to a Pareto or trend analysis. These tools are of course statistical tools.

I don't think I would readily accept a check sheet, in itself, as demonstration of use of statistical tools.

Dave

 

[Jim Wynne]

As it mentions in TS 16949 requirement
7.5.1.3 The organization shall use statistical methods of
verification where applicable.

and

8.1.1 Appropriate statistical tools for each process shall be determined during advance quality planning and included in
the control plan.

Could the word "statistical tools and method" be considered as a check sheet?

My opinion is "check sheet" can be considered as a basic statistical tools, since, we utilize the data (mean, standard deviation, ....) for decision making.

Thank for your comment

IMO, a checksheet becomes a "statistical tool" when results from it are used to support improvement. Any sort of data collection should qualify, so long as the data is used to do something.

 

[dokes]

I agree that nearly anything used to collect and/or analyze data could be considered a statistical tool. For example, a pareto diagram performs basically the same rough function as an ANOVA ... identifying what is most significant. A run chart certainly would also qualify, allowing analysis of trends. Histograms allow looking for non-normal distributions that can indicate outliers, multi-modes indicating multiple distributions (special causes), etc.

 

[qualitygoddess - 2010]

Checksheets can be formatted to provide a simple visual histogram of the data collected. See Ishiskawa's Guide to Quality Control. You'll see a simple example in there.

So, yes a checksheet is a quality tool and it can be used in a statistical analysis. It could be argued that it is therefore a statistical tool.

 

[sakkapak]

Oh I see.
That means it depends of the format of check sheets and utilization of them.

Thank you very much.

 

[Statistical Steven]

Statistical methods and statistical tools in general requires that tools such as ANOVA, DOE, regression, etc should be used when appropriate. Check sheets are data collection tools and the basis for statistical tools, but by themselves are NOT statistical tools or methods.

 

[Jen Kirley]

The operative words being "when appropriate." The more sophisticated and admittedly mathematically correct SPC methods wouldn't work in almost every application that I have seen--mostly small operations without the process refinements for SPC to work.

One can't run SPC when a process isn't contollable, though one can do things like run some P charts for improvement purposes. A check sheet may also satisfy the element's purpose if it is appropriate and not just an exersize to say, "See here, we are doing SPC."

 

[Jim Wynne]

Statistical methods and statistical tools in general requires that tools such as ANOVA, DOE, regression, etc should be used when appropriate. Check sheets are data collection tools and the basis for statistical tools, but by themselves are NOT statistical tools or methods.

"DOE" is not a statistical tool any more than architecture is a carpentry tool. Experimentation is, but design of experiments isn't. You cite a couple of different types of analyses, and of course there are many more, but data collection and analysis is use of "statistical methods" no matter how you look at it. There is nothing that "requires" use of those or other sophisticated methods in order to qualify as statistical methods in the ISO standard.

 

[Statistical Steven]

"DOE" is not a statistical tool any more than architecture is a carpentry tool. Experimentation is, but design of experiments isn't. You cite a couple of different types of analyses, and of course there are many more, but data collection and analysis is use of "statistical methods" no matter how you look at it. There is nothing that "requires" use of those or other sophisticated methods in order to qualify as statistical methods in the ISO standard.

I guess this is why I do not audit, and stick to being a statistician. DOE to me is a statistical tool, because it ensure experimentation without bias. Nonetheless, I believe that the intention of ISO and other standards that require appropriate statistical methods is not just to use statistics, but to use them correctly.
I come from the Deming school of quality. I believe each organization should have a "Master" Statistician that is responsible for ensuring the correct methods are used, and training all members of the company to use statistics.