[Ifeffit] Chi in arthemis

[Matt Newville]

Dear Eugenio,

> I see that the R-factor is pretty good, 1.74%, amp is high cause is
> correlated with the coordination number and always have big errors., delr
> has the error of the total distance, so it is ok, but ss and enot have a a
> really big error, is this normal?

What leads you to conclude that the uncertainty in delr (of ~0.016
Ang) is OK, but the uncertainty in ss (~0.0025 Ang^2) and enot (~2 eV)
are really big?  I don't know how to interpret the value for "amp" or
assess why it would have a "big error" (is that a big error?  how  is
"amp" used in your model?).

Generally, for "decent data" and a good fit, R has uncertainty of 0.01
or 0.02 Ang, E0 has uncertainty of 0.5 to 1.0 eV, and the amplitude
factors (N and sigma^2) are good to 10% or so.  Without knowing more
details, I'd say that your results fit within the "normal" range.

Having a "reasonable R-factor" of a few percent misfit and a reduced
chi-square of  ~100 means the misfit is much larger than the estimated
uncertainty in the data.  This is not at all unusual.   It does not
necessarily  mean (as Scott implies) that this is because the
uncertainty in data is unreasonably low, but can also mean that there
are systematic problems with the FEFF calculations that do not account
for the data as accurately as it can be measured.   For most "real"
data, it is likely that both errors FEFF and a slightly low estimate
for the uncertainty in the data contribute to making reduced
chi-square much larger than 1.

And, yes, the community-endorsed recommendation is to report either
chi-square or reduced chi-square as well as an R-factor.  I think some
referees might find it a little deceptive to report  R-factor because
it is "acceptably small" but not reduced chi-square because it is "too
big".

--Matt

On Tue, Aug 18, 2009 at 5:18 PM, Eugenio Otal<eugenioh at gmail.com> wrote:
> Hi Scott,
> here I copy a part of the report:
>
> Independent points          =       6.222656250
> Number of variables         =       4.000000000
> Chi-square                  =     247.145092496
> Reduced Chi-square          =     111.193574128
> R-factor                    =       0.017422216
>
> Guess parameters +/- uncertainties  (initial guess):
>   amp             =     6.7815290   +/-      1.4687660    (1.0000)
>   enot            =     2.2173620   +/-      2.1499920    (0.0000)
>   delr            =     0.0514640   +/-      0.0163900    (0.0000)
>   ss              =     0.0074020   +/-      0.0025220    (0.0030)
>
> I see that the R-factor is pretty good, 1.74%, amp is high cause is
> correlated with the coordination number and always have big errors., delr
> has the error of the total distance, so it is ok, but ss and enot have a a
> really big error, is this normal?
> Thanks, euG
>
> _______________________________________________
> Ifeffit mailing list
> Ifeffit at millenia.cars.aps.anl.gov
> http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit
>
>