Linearity assessment

[#1]

Dear all,

I want to perform a regression for method comparison in laboratory using two types of data :

- Data from a first analyzer in X axis which were performed in 1 run in duplicate (2 values)

- Data from a new analyzer in Y axis which were performed in 2 run in duplicate (4 values)

With regression data, I want then to perform least squares regression, lack of fit test and outlier test. Can anybody help me on this point ?

Thank you for your help

[#2]

Are you following a protocol such as EP-09 by CLSI?

JMP supports Deming regression and Passing-Bablok regression in Analyze > Fity Y by X (ultimately the Bivariate platform). It also supports the Bland-Altman analysis in Analyze > Specialized > Matched Pairs.

See the Bivariate and Matched Pairs documentation.

[#3]

Dear Mark, 

Thank you for your help.

Are there any tools for helping in data table preparation ? Especially for study including replicates.

Thank you,

Best regards.

[#4]

What is the question regarding replicates? Replicates appear as additional rows.

[#5]

I should also mention that the Measurement System Analysis platform offers a lot of capability for comparisons, including bias and linearity. See the MSA documentation.