Hi,

Is there a an Excel Template that handles the Attrbute Gage R&R with ORDINAL Data (i.e. 1,2,3,4). I've searched around but I can only find the an excel template that deals with BINARY Data (i.e. Pass/Fail). I hate to start from scratch if there is already a file floating around. Thanks.

Manny

Hi,

Is there a an Excel Template that handles the Attrbute Gage R&R with ORDINAL Data (i.e. 1,2,3,4). I've searched around but I can only find the an excel template that deals with BINARY Data (i.e. Pass/Fail). I hate to start from scratch if there is already a file floating around. Thanks.

Manny

Welcome to the Cove, Manny. When in doubt, do a search--there aren't too many issues that we haven't discussed, and there are a lot of file attachments. For example, have a look at the attachment in this discussion thread: Gage R&R Studies for Attribute Data (http://elsmar.com/Forums/showthread.php?t=18734).

Thanks Jim. I pretty much look at all the files in the free directory (http://elsmar.com/pdf_files/) and forums (http://elsmar.com/Forums/fileslist.php) here but to no avail. I guess I just have to customized my own based on one of the posted files which are all based on BINARY data. Again, I need a template for ORDINAL DATA, same concept but I have to redo all the formulas and even the format.

:thanx:
Manny

Thanks Jim. I pretty much look at all the files in the free directory (http://elsmar.com/pdf_files/) and forums (http://elsmar.com/Forums/fileslist.php) here but to no avail. I guess I just have to customized my own based on one of the posted files which are all based on BINARY data. Again, I need a template for ORDINAL DATA, same concept but I have to redo all the formulas and even the format.

:thanx:
Manny

Maybe if you could explain a little about what you're trying to do, someone here can help. By definition, "attribute" data is binary--good/bad, pass/fail, etc., so what sort of measurement data are you dealing with?

you need a KAPPA test. it allows for more 'bins' than just pass or fail. so your ordinal data will work.

I have attached a spreadsheet. you will find it in a couple of other posts I've made if you search for Kappa. you will also find a link to a quality progress article that explains the situation fairly well.

you need a KAPPA test. it allows for more 'bins' than just pass or fail. so your ordinal data will work.

I have attached a spreadsheet. you will find it in a couple of other posts I've made if you search for Kappa. you will also find a link to a quality progress article that explains the situation fairly well.

the number of categories = the number of ordinal categories...

you need a KAPPA test. it allows for more 'bins' than just pass or fail. so your ordinal data will work.

I have attached a spreadsheet. you will find it in a couple of other posts I've made if you search for Kappa. you will also find a link to a quality progress article that explains the situation fairly well.

the number of categories = the number of ordinal categories...

Thanks Bev D. Very neat. Interpretation Tab in the same work sheet will be a nice addition.

I can see the Comparison between sample Vs categories.
Where do you input individual Judges Scores?

I am attaching a hyperlink to Ordinal data Attribute R & R using Minitab Attribute agreement analysis.
Here is an example with that application
www.minitab.com/resources/tutorials/AttributeAgreementAnalysis/SQAttributeAgreementAnalysis.pdf
Regards,
Govind.

you need a KAPPA test. it allows for more 'bins' than just pass or fail. so your ordinal data will work.

I have attached a spreadsheet. you will find it in a couple of other posts I've made if you search for Kappa. you will also find a link to a quality progress article that explains the situation fairly well.

the number of categories = the number of ordinal categories...

Bev, I linked to a post in which you attached that file, and the OP said it's not what he's looking for, which was why I then asked him to to explain further, because it appeared to me that your Kappa Excel sheet should have been just the ticket.

Thanks Bev D. Very neat. Interpretation Tab in the same work sheet will be a nice addition.

I can see the Comparison between sample Vs categories.
Where do you input individual Judges Scores?


just ener the number of samples, categories adn judges. then click on the "make table" button. The rows in the table are the sample number and the columns are the categories. the number of judges who score a specific sample as belonging to a specific category (or having a specific value) are entered into the corresponding cell...try it out.
I have further attached an example I use to introduce the concept to my students. I show them 11 different colors and give them only 6 color categories: red, blue, purple, green yellow and white. then for each of the 11 colors I show them (#sampels = 11, number of categories = 6) I enter teh number of 'votes' for each color. each row must add up to the total number of judges, in the example I attached, there are 16 judges.

Bev, I linked to a post in which you attached that file, and the OP said it's not what he's looking for, which was why I then asked him to to explain further, because it appeared to me that your Kappa Excel sheet should have been just the ticket.

well I guess we need to hear from the OP then, as the spreadsheet will work for ordinal data. I use it for that quite a bit.

I see a lot of people suggesting the use of a Kappa test as a substitute for an R&R for ordinal data, but this is not correct. Kappa is ok for ATTRIBUTE data where the attributes are not ordered. However, if you have ORDINAL data, the test needs to take into account that not all disagreements are equal. For example, if you have operators grading someting on a [1,2,3,4,5] scale, it's important to consider that a pair of grades [1,2] have more agreement than a pair [1,5].

What you really want to use is Kendall's Coefficient of Concordance. Minitab has a very good tutorial on this... but I can't post the link so I'll just suggest you google "Minitab Attribue Agreement Analysis Example".

To do this with Minitab 15, you can use the tools under Stat->Quality Tools->Gage Study to generate a blank table and then analyze the data.

Sadly, I do now know of a free Excel template for this... but maybe an enterprising individual will make us one. You can also buy an excel package to do this sort of thing... if you search for "Kendall Tau Correlation Excel" you will find a few different options.

Cheers
-Bill