Spike-and-recovery and linearity

[montana]

Hi,
I would to ask you about validation of linearity , i am using a spike and no spike samples. I have two linearity curve, how can i compare the two slopes (linear regression)=statistic test.

The standards:
y = 1348,x - 450,3
R² = 0,999
Plasma spiked-Plasma blank
y = 1346,x + 246,9
R² = 0,993

[tom jupille]

You need the standard error (SE) of one of the the slopes (which Excel will calculate as part of its regression calculation). The 95% confidence interval is (*very* roughly) twice the standard error. If the second slope is within +/- 2*SE of the first, then there is only a 5% probability that they are different.

For a *much* more thorough explanation, check the first five installments of Coleman & Vanetta's "Statistics in Analytical Chemistry" series in American Laboratory:
http://www.americanlaboratory.com/841-S ... 0chemistry