Dear all,

For bias in MSA 3rd edition, I have one question as following.
When calculate the 1-α confident interval (CI), there is one formula:

Bias+/-[d2σb(tv,1-α/2)/d2star]

The question is why d2/d2star is used, because when I check the traditional t distribute CI formula, there is no such kind of parameter.

Your attention would be highly appreciated.
Best regards
Steven

Any experts here have an answer for this post?

It's a weekend. Many of our regular Covers take the weekend off and will return Monday. I'm sure some of the statistics whizzes will have an interesting answer.

hi
in AIAG MSA , use d2 and d2* to calculate the range of bias ,rather than use T-test to test the bias
thanks

For bias in MSA 3rd edition, I have one question as following.
When calculate the 1-α confident interval (CI), there is one formula:

Bias+/-[d2σb(tv,1-α/2)/d2star]

The question is why d2/d2star is used, because when I check the traditional t distribute CI formula, there is no such kind of parameter.
hi
in AIAG MSA , use d2 and d2* to calculate the range of bias ,rather than use T-test to test the bias
thanks
My knowledge of statistics is rather limited so I won't be able to explain this esoteric stuff well. From whatever I've tried to research on the subject, here are a few links that MAY help.

From AIAG: StatExpose_d2.pdf (http://www.aiag.org/forms/StatExpose_d2.pdf)

From Quality Engineering, Vol 13, No 2: (Word Doc)
A Note on Using Ranges to Estimate Variability (http://ceaspub.eas.asu.edu/dcm/A Note on Using Ranges to Estimate Variability, Published in Quality Engineering, Vol 13, No 2, pp. 211-217.doc) by Woodall/Montgomery.

From iSixSigma Discussion forum: Bias Calculation (http://www.isixsigma.com/forum/showmessage.asp?messageID=33854)

and

http://elsmar.com/Forums/showthread.php?t=7178

almost two years passed, we have no anwser,I think we did better put forwards this to AIAG, maybe it is mistake ,as i know they made mistake in other place,they mistake “g” for “ mg”.

Atul Khandekar ,thank you for your offer , I get answer now, in fact ,this is my question, I ask steven Li put forward it