Interpretation of Shapiro--Wilks results
Hiya. I've got some data that I'm trying to fit a normal curve to; S/
W gives me a W very close to 1 (that's good -- indicates normality)
but the significance is 0, apparently meaning that, as it's <.05, I've
still got to reject the null-hyp of normality, even though W is high.
I'm trying to understand how to interpret a high W with a p of 0 ...
and all of the plots (P-P, Q-Q, Detrended Q-Q, boxplot) support
normality, too.
Drop me a note, please (doug.elias@xxxxxxxxx), I don't follow this
group very closely.
Thanks,
doug
.
Relevant Pages
- Interpretation of Shapiro--Wilks results
... I've got some data that I'm trying to fit a normal curve to; ... still got to reject the null-hyp of normality, ... and all of the plots (P-P, Q-Q, Detrended Q-Q, boxplot) support ...
(sci.stat.edu) - Re: Interpretation of Shapiro--Wilks results
... but the significance is 0, apparently meaning that, as it's <.05, I've ... still got to reject the null-hyp of normality, ... and all of the plots (P-P, Q-Q, Detrended Q-Q, boxplot) support ... post in the "Is there a mininum sample size for normality tests?" ...
(sci.stat.consult) - Re: Normal curves
... Superpose a normal curve with these parameters and then provide the ... this is the case and you use normality as your criterion for data ... Contacting a statistical consultant is the most remote possibility. ...
(sci.stat.math) - Re: Skewed differences and paired students t test
... > Thanks Dave & Rich for the detailed replies here. ... > page where someone used them to measure normality of data, ... > beta-binomial model would be the best fit. ... closer to Normal than the original populations. ...
(sci.stat.math) - Re: Skewed differences and paired students t test
... in my data that I'm trying to fix. ... someone used them to measure normality of data, ... a power transformation to improve the fit to normal. ... "fairly" normal without noticable skew. ...
(sci.stat.math) |
|
