Regression analysis with errors in both X and Y

[FairProductFox3]

There should be an add-in which allows generating regression lines accounting for errors in both X and Y. Usually york fit is used for this purpose in other software. It is crucial that JMP also introduces this feature.

[jszarka]

Starting in JMP 17, you can use the Passing-Bablok fit: Passing-Bablok Fit Report (jmp.com)

In the Bivariate platform, use the Fit Passing Bablok option to fit a regression model using the Passing-Bablok method. Passing-Bablok regression was developed for comparing measurements from two different analytical methods. The method assumes a linear relationship between the two variables with strong correlation. A line of fit and a dotted line for Y = X are added to the scatterplot. Use the Bland Altman Analysis option in the Passing-Bablok Fit red triangle menu to perform a paired t-test and Bland-Altman analysis.

[hogi]

If you can specify the variance ratio between both errors, you can use Deming fit, which is available via:

The actual issue in JMP:
the best regression options are hidden in the bivariate Report - and not available via Graph Builder !??!
here is a wish to fix the issue: Graph Builder: robust regressions 

[SarahGilyard]

As mentioned in a previous comment, you can perform a Deming regression in JMP under the "Fit Orthogonal" menu and specify the error ratio between X and Y. This assumes a constant measurement error ratio between the variables and expects that errors are normally distributed and independent. The Passing-Bablok might be the better choice which has less requirements for the error distributions of X and Y. As mentioned in a previous comment, this regression option is now available. 

The York regression is an interesting suggestion. We will consider the expanded use-cases of this vs. the Deming and Passing-Bablok options currently available. Thank you for posting. 

[hogi]

The robust regressions are sooo useful.
On the other hand: most users use Graph builder.

Is there any chance to provide this functionality to the majority of the users:
Builder: robust regressions