How to understand (analyze) blind duplicates/replicates
Thread Content
This post was last edited by echommm on 2010-4-28 at 10:26. ASTM D6378-08 – Method for Determining Vapor Pressure of Petroleum Products, Hydrocarbons, and Hydrocarbon-Oxide Mixtures. X3.2 Samples: X3.2.1 The researchers analyzed randomly arranged sets of samples, which consisted of blind duplicates and replicates taken from the same containers holding 20 different types of hydrocarbons and hydrocarbon-oxide blends, stored in 250-mL and 1-L clear glass containers. In other words, there were 20 samples per type, with 2 duplicates each, across 2 container sizes, and 2 replicates per bottle – resulting in a total of 160 analyses/data sets. The 20 sample types included 6 conventional gasolines (with and without ethanol added), 5 reformulated gasolines containing either MTBE or ethanol, 6 CARB gasolines (2 without oxygenates, 2 with MTBE, and 2 with ethanol), 1 aviation gasoline, 1 JP4 fuel, and 1 condensate. The total vapor pressure of these samples ranged from approximately 22 to 110 kPa (3.2 to 16.0 psi). :dizzy: :dizzy:
-- Does \"blind duplicate\" refer to an empty backup? “Replicate” should mean to copy, right? --
The latter formula is interesting: 20 samples X 2 duplicates X 2 bottle sizes X 2 replicates/bottle = 160 analyses/data sets. It can be seen that the roles of \"duplicate\" and \"replicate\" are different; I guess 1 sample = 2 duplicates, 2 bottle sizes = 4, and replicates = 8
For each sample, (X---large bottle, x---small bottle) X----------- XY-----------------XxYy---------------------XXxxYYyy duplicate 2 bottle size replicates/bottle 1-------------2 ----------------- 4 --------------------------8 20*8=160
This post was last edited by niuky on 2010-4-28 at 16:53. “Blind” seems to mean unknown; I guess it refers to the fact that the analysis laboratory does not know the origin and properties of the sample. Duplicates refer to taking two samples from the same sample; for example, two bottles from each of the large and small bottles in this case. Take sampling replicates. Replicates refer to the fact that two parallel samples are tested for each vial of sample during analysis. Analyze the parallel samples.
I think Niuky’s analysis makes a lot of sense. “Blind” refers to something that is unknown or undetermined; I don’t understand what the purpose of using large and small test samples is… “Duplicate” refers to the larger twin test samples needed for comparison, while “replicate” refers to the smaller parallel samples generated from them. I’m wondering if it would be possible to use a master sample and sub-samples……
Is the difference between duplicates/replicates a quantitative one? The former is two, while the latter can be multiple? I haven’t figured it out yet; I’ll go consult a professional analyst today (of course, good English skills are required, otherwise it’s impossible to understand). I will post the results once they are available. Thank you all for your enthusiastic participation! :handshake :handshake
duplicate: parallel sample; replicate: repeated sample
7# echommm From the original formula, \"duplicate\" refers to repeated sampling, while \"replicate\" refers to repeated analysis; it seems to have nothing to do with the number of times