P Value Usage
Six Sigma – iSixSigma › Forums › Old Forums › General › P Value Usage
 This topic has 3 replies, 3 voices, and was last updated 14 years, 6 months ago by watkins.

AuthorPosts

November 19, 2007 at 9:09 pm #48717
Hi,
I am new to six sigma and am trying to understand a couple of things.
1) Why do we use the p value to determine if data is stable etc ? Why Specifically 0.05 why not 0.1 ?
2) What is the difference between Percentages and Percentiles.. Is Span a good metric to look at while doing projects ?
Rgds,
Tom.0November 19, 2007 at 9:14 pm #165126My quick response is we don’t use pvalues to assess the stability of a process, we use control charts to help distinguish random from nonrandom variation..05 is a default value, but you should not be “blinded” by it. It is completely appropriate to change that to .10 or .01 in some cases. It helps you assess risk of over reacting (rejecting the null when you shouldn’t have).Span (Range, Standard deviation, Variance and Sum of Squares) are all ways to describe variability in the data (process). Welcome to the Discussion Forum
0November 19, 2007 at 9:26 pm #165133Thanks for the quick reply Pete. I surely want to start learning more and what better place than isixsigma.
I beleive when i said p values for Stability i was referring to the fact that we use the p value for clusters, mixtures, trends and oscillations to figure out if the data is stable ..we also look at P value for normality . How did the whole 0.05 concept come into place ? I would love to get into training one day and i am sure its something i as a student would definitely ask.
Also, am still trying to understand the difference between percentile and percentages ?
Rgds,
Tom0November 19, 2007 at 9:42 pm #165143The concept of 0.05 refers to alpha risk, which is a pretty fundamental tenet of hypothesis testing. For any test, we state in advance how willing we are to tolerate the chance that we’ll end up rejecting the null hypothesis when we shouldn’t. Generally speaking, classroom settings advocate an alpha risk of five percent, simply because of existing practice and norm. However, pharmaceutical companies bringing a drug to market might have to conduct tests using an alpha of 1 percent, while a lowrisk business process might opt for 10 percent. It simply depends.
A percentile is a threshold or level in a series of ordered observations, while a percentage is simply a fraction or proportion of the series with a particular characteristic. For example, out of 500 graduating students, students in the 95% percentile are those students in the top five percent – or the top 25 – of all 500 students based on final grade. As another example, if 200 of these students are boys, then boys make up 40% of the class.0 
AuthorPosts
The forum ‘General’ is closed to new topics and replies.