Chem 8157

I’m assuming that the value reported for intralaboratory RSD is an average of the intralaboratory RSD values from each lab, since the values reported in the table are averages of the samples. It would have been clearer had the authors included an RSD or the standard deviation of samples from each lab.

for 1+10, does lab A have an intralab RSD based on their data, and so does lab B, C, D.... and the final intralab RSD is the average?

I agree with Melissa, however I am confused by the fact that there are 8 different labs used, but only one intralaboratory precision value given. In my opinion, this would seem more like an interlaboratory precision value, as several laboratories are in use.

I found some knowledge about intralaboratory precision from google results. Intralaboratory precision is
calculated using Youden pair, which are two samples with similar
concentration. By drawing the Youden plot for each sample pair at
different labs within one graph, one will get information related to intralaboratory
precision. The intralaboratory precision is the square root of variance of D (the difference of the two samples). So it requires the results form all the laboratory to give all the D data. It means the intralaboratory
precision represents the data of all the labs. So in the table, for
each sample combinatioin at different labs, one only get one
intralaboratory precision.
reference:
http://books.google.com/books?id=SJP57-0UOzYC&pg=PA236&lpg=PA236&dq=youden+intralaboratory+precision&source=web&ots=HdPFKdnvVa&sig=LsWVGaC2-tbKoVdZOMOq0gtuZq8
http://books.google.com/books?id=9PD96gsJtk8C&pg=PA57&lpg=PA57&dq=youden+intralaboratory+precision&source=web&ots=tL0n4L7BRh&sig=Durw5WzRIPNOqglsOzM0bJxohLw

I've read Jing Zhang's reference. It's perfect to answer this question. I totally agree with it.

Thanks for Zhang Jing's reference. But I still confused about the difference between Intralaboratory and Interlaboratory precision. So I add the definition of Intralaboratory and Interlaboratory precision to know the difference more easily.
Intralaboratory precision reflects the ability of trained laboratory personnel to obtain consistent results repeatedly when performing the same test.
Interlaboratory precision is a measure of reproducibility of a method when tests are conducted by a number of laboratories using that method. Generally, intralaboratory results are less variable than interlaboratory results.

SUMMARY AND COMMENTS.

The best way to address questions like this one is to refer to the original reference give in the example (i.e. Journal of AOAC International 82, 1999, 112-118). This probably means a light stroll to Walter Library.

In the event that the reference is not available, one can identify a published method that could be applicable.

1. The Youden Pair Plot seems applicable. It is graphical and needs that the participating labs have the same concentrations analyzed twice. The difference between their results of each lab is a measurement of their precision. Please note that it is the square root of the variance what is being used, not the average.

2. When two samples are not analyzed, one needs to use multifactorial analysis. An ANOVA test would be suitable. This test will be able to distinguish errors associated with several factors (e.g. different laboratories, procedures within the same laboratory, sample preparation, etc.). The caveat is that the experiment needs to be planned keeping in mind that ANOVA will be used.