hi, actually I am working on comparing several analytical validation method
what i expect that in the classical approach of analytical approach is obligatory that the variance must be homogeneous
my question about the new approach of Accuracy profile the variance should be also homogeneous or not
Yassine -
I have done a great deal of work involving heteroscedasticity in a different context, and I am not sure if that is what you mean, but heteroscedasticity is a very natural part of the error structure in many various regression applications. Or are you comparing various models with different homoscedastic variances?? I am not familiar with the term "accuracy profile," but a quick check seemed to indicate it is a term from analytical chemistry. ??
At any rate, I think you might, at least peripherally, be interested in the following paper by retired analytical chemist, Christopher Lee, involving heteroscedasticity:
https://www.researchgate.net/publication/262524025_Calibration_uncertainty_in_pharmaceutical_analysis_replicate_single-point_calibration_revisited
If this is not related to your question, I think it still might possibly be of interest to you.
Cheers - Jim
PS - Ah. I assumed you were looking at regression, but maybe not? Perhaps you mean something more along the lines of tests noted in the following question responses:
https://www.researchgate.net/post/How_to_analyze_data_with_heterogeneity_of_variance2
If you look at any hypothesis tests, remember that a lone p-value is not enough:
Press release for the American Statistical Association:
http://www.amstat.org/newsroom/pressreleases/P-ValueStatement.pdf
--------------------
https://www.researchgate.net/publication/262971440_Practical_Interpretation_of_Hypothesis_Tests_-_letter_to_the_editor_-_TAS
Article Calibration uncertainty in pharmaceutical analysis: replicat...
Article Practical Interpretation of Hypothesis Tests - letter to the...
thank you sir for your fruitful answer
lastly I am working with my laboratory about project of improving new analytical validation method, they are new method basing on statistics and applied on analytical chemistry in this way.
The classical method impose that the variance of all series of measurement should be homogeneous before following other steps in analytical validation this parameter is checked with the Cochran Test.
But the new way of analytical validation is checked with accuracy profile.
Accuracy profile is new graphical tool decision based on the approach of total error.
It is assumed that any analytical result should be closely to the true value that why the two important parameters of error which are systematic and random error are the principal component of this difference between the measured value and the true one.
the estimation of the tolerance intervals (limits) which gathering this total error and which give the expectation value for the routine application is based on beta expectation tolerance interval and beta content tolerance interval
I wish that i give more explanation about our work
Validation is good, but sorry, I don't know the technique to which you refer. - Best wishes with your endeavors.
Dear Yassine,
Maybe my answer is not in this direction, but recently we publish an article comparing two analytical tools for diagnostic tests, but could be expanded to other toolswe also provide a macro in Microsoft Excel to calculate the accuracy, maybe this can help you with something!
DOI: 10.1039/C7AY00505A
https://www.researchgate.net/publication/314560508_Improved_assessment_of_accuracy_and_performance_indicators_in_paper-based_ELISA
https://www.researchgate.net/publication/315493970_m-Accuracy_Macro_in_Microsoft_ExcelR_to_estimate_test%27s_accuracy
Article Improved assessment of accuracy and performance indicators i...
Data (m-Accuracy): Macro in Microsoft Excel® to estimate test’s accuracy