question about confidence intervals in HPLC assay

[usv]

I am fairly new to this so please excuse my ignorance.
We are looking at HPLC/UV assay results obtained on different instruments and days. I want to figure out the day to day variance of the method. Why are 3 of my actual values outside the range of 0.960 to 0.968 (data set example below)? Does this not seem high?

Also in looking at condifence interval for an assay calculation - how do you factor in any variance related to the assay std? The assay result is calcualted fromthe ratio of the avg of the sample injections to the assay std injections on a give day.
I appreciate your advice. Thanks.

Day Sample assay result
1 0.958
2 0.964
3 0.959
4 0.966
5 0.963
6 0.973
Average 0.964
stdev 0.005
%RSD 0.54
confidence 0.004
Range for 95% CI 0.964±0.004

[tom jupille]

Why are 3 of my actual values outside the range of 0.960 to 0.968 (data set example below)? Does this not seem high?

I suspect you're misinterpreting "confidence interval". The 95% confidence interval doesn't imply that 95% of your measurements will lie in that range, it implies a 95% probability that the true mean value is within that range.

how do you factor in any variance related to the assay std?

For independent errors, variances are additive. So each of the "values" in your data set is a mean of several injections and has its own variance:

σ² = σ²(sample) + σ²(standard)

[aceto_81]

Also to find the between day variance, you should measure at different days. From this you can find the between day variance and the within day variance.

Ace

[usv]

Thank you both for your suggestions. I need a little more help with the speciics of determining hte variance for an assay using the
σ² = σ²(sample) + σ²(standard)

If I am calculating variance for a given day's assay. I have an average area count and std deviation for the sample and average area count and std dev assay std. I can sum the square of the std deviations but I am not sure how to report it realtion to the assay value since that is the ratio of the area sample to area std.

e.g.
Sample Std
39297.060 41005.3 average area
316.976 145.879 daily stdev
100473.688 21280.788 daily sigma squared for sample and std
sum of sigma squared = 121754.476
calculated assay = 0.958 How would I report the variance including the variance for both the sample and std in the assay (ratio) result?

Thank you for your help.

[aceto_81]

e.g.
Sample Std
39297.060 41005.3 average area
316.976 145.879 daily stdev
100473.688 21280.788 daily sigma squared for sample and std
sum of sigma squared = 121754.476
calculated assay = 0.958 How would I report the variance including the variance for both the sample and std in the assay (ratio) result?

Thank you for your help.

If your sample is 1 sample, but multiple sample preps, and your std is 1 standard preparation, then you can state:
The total variance due to injection and sample preparation = 100473.688
But the variance by the injection = 21280.788
So the variance caused by the sample prep = 100473.688-21280.788
= 79192.9
Or sqrt(79192.9) = stdev = 281.4 = 281.4/39297.060*100 = 0.72%


Ace