# normality

Viewing 8 posts - 1 through 8 (of 8 total)
• Author
Posts
• #47812

Katrina
Participant

hey, what is the reasoning for the a value >0.05 to signify normality?! Why has this value been chosen as the cut off point?!
thanks
kat

0
#159861

Aguilar
Member

http://www.sportsci.org/resource/stats/pvalues.html
Cheers

0
#159865

Katrina
Participant

cheers kat

0
#159873

Katrina
Participant

hey, Ive read that article and im still not clear as to why a >0.05 value would signify that the data is normal?! why this value?!
thanks
kat

0
#159880

Jim Shelor
Participant

Katrina,
The value 0.05 comes from the confidence level you want for your answer.  If you want a confidence level of 95%, then if P<0.05, you reject the null hypothesis because the p is significant.
In this case, the null hypothesis is the distribution is normal.  Therefore, a p<0.05 says to reject the null.  Accordingly, the distribution is not normal.
If for this case, null hypothesis = distribution normal, your p>0.05, there is insufficient evidence to show the null is not true at the 95% confidence level and you fail to reject the null hypothesis.  Your distribution is normal.
You can use a 90% confidence level, in which case the significant p level is 0.1; or a 99% confidence level in which case your significant p level is 0.01 (1-confidence).
I hope this helps.
Respects,
Jim Shelor

0
#159885

BTDT
Participant

Kat:The number is based on this statement from R.A. Fisher (1890-1962) on whether something that occurs is due to random chance or otherwise. It was pretty arbitrary, but everyone has come to agree to the 1 in 20 rule.… either there is something in the treatment, or a coincidence has occurred such as does not occur in 1 in 20 trials (Journal of the Ministry of Agriculture of Great Britain, 1926)Cheers, BTDT

0
#159905

Katrina
Participant

thanks for your help guys, think i get it now, was over complicating things lol!
cheers
kat

0
#159907