-
- admin
- Site Admin
-
- Posts: 647
- Joined: Wed Aug 11, 2004 12:19 am
By Roel Merckx on Thursday, August 19, 2004 - 12:24 pm:
What are the possible causes or how to interpret a calibration curve which doesn't pass through zero?
-------------------------------------------------------------------------------------------------------
By bert on Thursday, August 19, 2004 - 11:18 pm:
What do you mean by not passing through zero? Did you include a blank, which revealed response for your analyte or does the calculated regression line not pass zero. If it is the regression line, is the deviation from zero statistical significant?
-------------------------------------------------------------------------------------------------------
By Brent on Friday, August 20, 2004 - 07:17 am:
It doesn't have to pass through zero, though often the y-intercept is small enough that it might as well be zero.
Here's a nice little test. Integrate a blank sample with no peak. You will not get zero integration.
-------------------------------------------------------------------------------------------------------
By syx_gf on Sunday, August 22, 2004 - 05:48 pm:
there is a reference from vam about preparation of calibration curve: a guide to best practice.
It said that leverage problem of a calib curve can be arose when calib stds are prepared by sequential dilution of a stock soln or in other words, the stds should be independent, and should not be prepared from a common stock soln)
What are the possible causes or how to interpret a calibration curve which doesn't pass through zero?
-------------------------------------------------------------------------------------------------------
By bert on Thursday, August 19, 2004 - 11:18 pm:
What do you mean by not passing through zero? Did you include a blank, which revealed response for your analyte or does the calculated regression line not pass zero. If it is the regression line, is the deviation from zero statistical significant?
-------------------------------------------------------------------------------------------------------
By Brent on Friday, August 20, 2004 - 07:17 am:
It doesn't have to pass through zero, though often the y-intercept is small enough that it might as well be zero.
Here's a nice little test. Integrate a blank sample with no peak. You will not get zero integration.
-------------------------------------------------------------------------------------------------------
By syx_gf on Sunday, August 22, 2004 - 05:48 pm:
there is a reference from vam about preparation of calibration curve: a guide to best practice.
It said that leverage problem of a calib curve can be arose when calib stds are prepared by sequential dilution of a stock soln or in other words, the stds should be independent, and should not be prepared from a common stock soln)
