Received: from imo-d23.mx.aol.com(205.188.139.137) by zeta.aps.anl.gov via csmap (V6.0) id srcAAATTa4cr; Sat, 28 Feb 04 13:35:23 -0600 Received: from SCalvin100@aol.com by imo-d23.mx.aol.com (mail_out_v36_r4.14.) id r.7e.481c3cf9 (18251) for ; Sat, 28 Feb 2004 14:35:19 -0500 (EST) From: SCalvin100@aol.com Message-ID: <7e.481c3cf9.2d724777@aol.com> Date: Sat, 28 Feb 2004 14:35:19 EST Subject: =?ISO-8859-1?Q?Re:=20[Ifeffit]=20question=20about=20S02=20(passi?= =?ISO-8859-1?Q?ve=20electron=20reduction=A0=20factor)?= To: ifeffit@millenia.cars.aps.anl.gov MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="part1_7e.481c3cf9.2d724777_boundary" X-Mailer: Thunderbird - Mac OS X sub 24 --part1_7e.481c3cf9.2d724777_boundary Content-Type: text/plain; charset="ISO-8859-1" Content-Transfer-Encoding: quoted-printable Matt, Good point. It is much more straightforward to check the uncertainty than th= e=20 correlation to determine if that could be responsible for the anomalous=20 S02...so I should have said that Yu-Chuan should check the uncertainty of th= e S02=20 parameter. If it is very large (say 0.8), then Ifeffit is in effect reportin= g=20 that it cannot determine S02 well for some reason. At that point, I think=20= it=20 is helpful to see if there is a high correlation, since that may give a clue= =20 as to the reason for the large uncertainty. So this brings up a (possibly contentious) point. In their reporting=20 recommendations, the IXS suggests reporting high correlations, particularly=20= when they=20 are between parameters that do not routinely show high correlations. What is= =20 the reason for this suggestion? I'm not criticizing it...just looking for th= e=20 rationale. --Scott Calvin Sarah Lawrence College In a message dated 2/27/04 5:27:23 PM, newville@cars.uchicago.edu writes: >=20 > But: the correlation between S02 and sigma2 is NOT the cause of a > value for S02 being low by a factor of 3.=A0 A high correlation means > that a reasonably good fit could also be found by changing both S02 > and sigma2 away from their best-fit value.=A0 This is not at all the > same as saying that either parameter has the wrong best-fit value. > In fact, the correlation between pair of variables says nothing > about the trustworthiness of the best-fit values, only how much the > uncertainties in the best-fit values depend on other variables.=A0 > And, just to be clear, the reported uncertainties already take these > correlation into account. >=20 --part1_7e.481c3cf9.2d724777_boundary Content-Type: text/html; charset="ISO-8859-1" Content-Transfer-Encoding: quoted-printable Matt,

Good point. It is much more straightforward to check the uncertainty than th= e correlation to determine if that could be responsible for the anomalous S0= 2...so I should have said that Yu-Chuan should check the uncertainty of the=20= S02 parameter. If it is very large (say 0.8), then Ifeffit is in effect repo= rting that it cannot determine S02  well for some reason. At that poin= t, I think it is helpful to see if there is a high correlation, since that m= ay give a clue as to the reason for the large uncertainty.

So this brings up a (possibly contentious) point. In their reporting recomme= ndations, the IXS suggests reporting high correlations, particularly when th= ey are between parameters that do not routinely show high correlations. What= is the reason for this suggestion? I'm not criticizing it...just looking fo= r the rationale.

--Scott Calvin
Sarah Lawrence College

In a message dated 2/27/04 5:27:23 PM, newville@cars.uchicago.edu writes:


But: the correlation between S02 and sigma2 is NOT the cause of a
value for S02 being low by a factor of 3.=A0 A high correlation means
that a reasonably good fit could also be found by changing both S02
and sigma2 away from their best-fit value.=A0 This is not at all the
same as saying that either parameter has the wrong best-fit value.
In fact, the correlation between pair of variables says nothing
about the trustworthiness of the best-fit values, only how much the
uncertainties in the best-fit values depend on other variables.=A0
And, just to be clear, the reported uncertainties already take these
correlation into account.


 --part1_7e.481c3cf9.2d724777_boundary--