I'll try to answer:
CV =sqrt(variance)/mean
It is not possible perform Student's T test because
only differences to the mean can be obtained and the
mean is not specified, althought Fisher-Snedecor's F
test can be performed:
Two samples from the same population (null Hy-
pothesis: same mean and variance)
sqr(s1)/n1 = sqr(2.77)* sqr(s2)/n2 or
sqr(CV1)=sqr(2.77)*sqr(CV2)
F Test value .95 =sqr(2.77)=6.77
F Test value .8 is tabulated
P. Belinchon
H. Provincial Badajoz
SPAIN
> -----Mensaje original-----
> De: [log in to unmask]
> [mailto:[log in to unmask]]En nombre de Samuel
> Vasikaran
> Enviado el: martes 5 de enero de 1999 10:15
> Para: [log in to unmask]
> Cc: [log in to unmask]
> Asunto: Critical Difference
>
>
> A change in an analyte of 2.77 times the CV would be considered
> significant with 95% confidence. What change is needed for 80%
> confidence?
> Can anyone enlighten me please?
>
> Sam Vasikaran
> Royal Perth Hospital
> Western Australia
>
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|