Accuracy???!!! - (Nov/03/2009 )

Hi all.. i ve got a crazy (may be even stupid) doubt all of a sudden (this happens when u analyse thousaunds of data )
I desire a certain amount A (i have added this component suppose its units is mg/ml)
I get the value from assay as B (from the graph equation again in mg/ml)
If i do B/A, the percentage of it is Accuracy... and 100-the answer is the inaccuracy!!!
If I do A/B, what does the answer signify???
Thanx.. hope i dint cinfuse!!!

suppose i did an ELISA
my standard was 2.5,5,10,25,50 ng/ml.
I plot a graph of the resopnse
i back calculate from the graph equation (linear or 4PL)
for 5 ng/ml standard, i get a value of 4 ng/ml
so the accuracy by which i could detect the standard (in other terms %recovery) is 4/5*100=80%. So the inacuracy is 20%
Now the inverse that is 5/4*100 is 125%. What does this value signify!!!
I hope its clear now!!!

To further simplify
value obtained/value desired*100. This gives acuracy
I wanted to ask what will value desired/value obtained give!!!
Both give different results... So the confusion!!!

Pradeep Iyer on Nov 4 2009, 02:49 AM said:

hi rick... long time

ya it actually describes the error % in the data.. so was just wondering if we can relate it to accuracy!!!
if i have no spike control.. and suppose its only the standard... wat term wud we give this value?? accuracy% (over or under-watever)?? or error% - bias in your terms!!!