Author Topic: PFE output file: %ERR is negative?  (Read 217 times)

GarnetSnowball

  • Post Doc
  • ***
  • Posts: 10
PFE output file: %ERR is negative?
« on: September 07, 2023, 06:38:26 PM »
Just a quick question.

The output file contains the column 'X %ERR' which, particularly when the mean value is low, can sometimes be negative.

My understanding is that % ERR is 1 standard deviation as a percentage, so why would this be negative?

Example file attached

Probeman

  • Emeritus
  • *****
  • Posts: 2771
  • Never sleeps...
    • John Donovan
Re: PFE output file: %ERR is negative?
« Reply #1 on: September 08, 2023, 08:07:15 AM »
Just a quick question.

The output file contains the column 'X %ERR' which, particularly when the mean value is low, can sometimes be negative.

My understanding is that % ERR is 1 standard deviation as a percentage, so why would this be negative?

Example file attached

When the k-ratio is negative the program maintains the negative result when the error is calculated.  Because you utilized the "Force Negative K-ratios to Zero" flag (under the Analytical | Analysis Options menu) you probably didn't notice this.

I do not recommend using the "Force Negative K-ratios to Zero" feature unless one is an expert and understands the implications of zeroing out data.  For example, bad background positions will produce over corrected net intensities, which for major elements, may not be recognized. Another example: when analyzing low concentration elements, the averages will be biased positive as results close to zero *should* have both positive and negative concentrations so they average to zero at zero.

Please turn this flag off unless you are an expert.
« Last Edit: September 08, 2023, 07:48:21 PM by Probeman »
The only stupid question is the one not asked!

GarnetSnowball

  • Post Doc
  • ***
  • Posts: 10
Re: PFE output file: %ERR is negative?
« Reply #2 on: January 12, 2024, 09:10:41 PM »
Thanks Probeman, this is very helpful