fed-up
22nd January 2009, 11:38 AM
In cusum charts when all the data points are on the target except for 1 or 2 of them that occur 1point above/ below the target. So even if its showing its significant. Is it statistically significant? Yes or No. And why?
Thanks
Steve Prevette
22nd January 2009, 01:11 PM
In cusum charts when all the data points are on the target except for 1 or 2 of them that occur 1point above/ below the target. So even if its showing its significant. Is it statistically significant? Yes or No. And why?
Thanks
I should point out that if you are referring to "target" as related to a specification, specifications have no bearing on whether or not something is statistically significant. There is a description of a "target" in Acheson Duncan's Quality Control and Industrial Statistics for a single sided Cusum. But just exceeding the target is not statisically significant even in this scheme.
In the two sided CUSUM chart, a special "mask" is made. If one or more points fall outside of the limits of the "mask", then you have a statistically significant trend.