I want to know limit of detection and quantification using JMP.

[#1]

I want to know limit of detection and quantification using JMP.

How should I treat data?

[#2]

In analytical chemistry, the limit of detection (LOD) is the smallest signal that is detectable with a given confidence.  This usually takes the form of a sample tested against a blank or reference standard.  In that case, you are talking about using a t-test with alpha equal to your confidence level.  This can be done in the Fit Y by X platform and also potentially in the Matched Pairs platform depending on the specifics of your experiment.

The limit of quantification (LOQ) is more arbitrary, but the rule that I was taught is 10x the standard deviation of the blank signal.  This can be determined using the distribution platform looking at the data from your reference or  blank.

Getting the LOD and LOQ  also depends on what form your data takes.  Can you help with some details about the data (number of measurements or maybe a sample data set)?  If you can the community might be able to better provide some more specific guidance.

Best,

M

[#3]

To add a bit to Mike's advice above you may find this information helpful:

Estimation of a Limit of Detection

[#4]

In Peters reference you will also be pointed to an article which talks about how to treat this kind of data:

JMPer-Cable: When-Responses-Are-Below-the-Limit-of-Detection

[#5 · Accepted solution]

I found this while searching, and it's a few years old. Please check out the new Limits of Detection features in JMP 16. There is now a column property, and it is used in both the DOE platform as well as in the Generalized Regression platform in JMP Pro. 

For Limits of Quantification, see Martin's response on this thread. 

Learn more:

Gain accuracy and precision for model estimates from limits of detection (LOD) control