excel function:
if the regression is : y-fit = a *ref + b
1/ s=steyx(bias,ref.)
2/ t(a)=abs(a)/(s/sqrt(devsq(ref.)))
p(a)=tdist(t(a),n-2,2)
t(b)=abs(b)/(sqrt(1/n+average(ref.)^2/devsq(ref.))*s)
p(b)=tdist(t(b),n-2,2)
3/ for individual part,
t(bias)=abs(average(bias))/(stdev(bias)/sqrt(n)) -----(n :the number of replicates)
p(bias)=tdist(t(bias),n-1,2)
for overall,
s for repeatbility=sqrt((s1^2+s2^2+...+s5^2)/5) ------(in this instance, there are 5 parts)
t(bias)=abs(average(bias))/(s/sqrt(n))
p(bias)=tdist(t(bias),n-1,2)
RSS Feed
33 个回复
huili81 (威望:0)
赞同来自: winkingyang 、我乐苦多
if the regression is : y-fit = a *ref + b
1/ s=steyx(bias,ref.)
2/ t(a)=abs(a)/(s/sqrt(devsq(ref.)))
p(a)=tdist(t(a),n-2,2)
t(b)=abs(b)/(sqrt(1/n+average(ref.)^2/devsq(ref.))*s)
p(b)=tdist(t(b),n-2,2)
3/ for individual part,
t(bias)=abs(average(bias))/(stdev(bias)/sqrt(n)) -----(n :the number of replicates)
p(bias)=tdist(t(bias),n-1,2)
for overall,
s for repeatbility=sqrt((s1^2+s2^2+...+s5^2)/5) ------(in this instance, there are 5 parts)
t(bias)=abs(average(bias))/(s/sqrt(n))
p(bias)=tdist(t(bias),n-1,2)