[Clascomment] OPT-IN: Measurement of two-photon exchange effect by comparing elastic e± p cross sections

[Daniel Carman]

Brian,

Thanks for the reply. I will be interested to see if any of the other readers shares my concerns about
how you describe your two data figures. I think the language that you use in the paper to describe the 
data reveals a bias in your consideration, namely that you have always felt that your data supports the 
size of the TPE effect that you believe explains the discrepancy. If this is not the case, then the language 
in the paper should have a few statements of your approach and how you sought to omit a bias. Certainly
you cannot claim a clear trend in Fig. 19 and claim there is no evidence for one in Fig. 20. Fig. 19 has a
single 1sig data point that is not unity and Fig. 20 has 3 points that indicate a possible Q2 dependence.

Regards,
Daniel

> On Feb 10, 2016, at 6:17 PM, Brian Raue <baraue at fiu.edu> wrote:
> 
> Hi Dan,
> Thank you for your comments, though the list is pretty short. You sure you did your usual careful reading? Responses...
> 
> General:
> - Use units of c=1 throughout as you are not consistent in your notation.
> Fixed, I think.
> 
> - You switch back and forth between "beam line" and "beamline". Be consistent.
> Fixed
> 
> Page 1:
> - Line 15. Use "... TPE effects has been observed."
> Done
> 
> Page 2:
> - Line 70. Use "... magnetic moment …".
> Done
> - Line 97. Use "... of the intermediate ...".
> Don't agree
> 
> Page 4:
> - Fig. 3. The representation of CLAS (the gray shadow) is far too dim and should
>   be made darker.
> When viewed on my screen it looks pretty good.  We'll see if the copy editors have an issue with it.
> - Line 305. Use "... torus magnet …".
> Well, technically there are 6 magnets. Plural form is appropriate.
> 
> Page 5:
> - Line 310. Use "... the spatial positions …".
> Unnecessary definite article.  Unchanged.
> - Line 317. Use "... downstream of CLAS.".
> Technically, we are saying "downstream of the CEBAF Large Acceptance Spectrometer" and the definite article is needed.  But I'll concede this since no "the" sounds better.
> Page 6:
> - Line 368. Use "... between the Moller and …".
> Definitely don't need that definite article.
> 
> Page 7:
> - Line 391. Use "Note that the distributions …".
> Oops. 
> - Line 393. Superscript problem on electron.
> Fixed
> - Line 407. Do not capitalize "Corrections".
> Fixed
> - Line 438. Use "... momentum corrections, …".
> Fixed
> - Line 448. Use "... with tracks of both positive ...".
> I think "events" is better than "tracks" here.
> 
> Page 8:
> - Line 484. Use "... as a $p$ candidate.".
> Okay
> - Line 486. Use "... identified as an elastic …".
> No. Here we identify THE elastic pair from among possible elastic pairs.
> - Line 571. Use "... for all but the …".
> You like your definite articles far more than I do.
> 
> Page 10:
> - Line 605. Use "... of sector 3 were removed ...".
> Nice catch. Done.
> 
> Page 11:
> - Line 670. Use "... energy distributions of the ...".
> Hmm, I suppose.  It really was a common energy distribution, though.
> 
> Page 12:
> - Line 734. "OPE" has not been defined in the paper.
> Yep.  And since this is the only place we used OPE I changed to a written out version
> - Line 734. Use "... between the OPE and TPE …".
> And I don't agree...
> - Comment on radiative corrections. Because of your tight kinematic cuts it seems that you
>   are also minimizing the radiative correction effects. This might be worth pointing out.
> In fact, we eliminate the pre-scatter radiative effects because that just changes the "beam energy." And it is true that the combined effect of the cuts is to essentially eliminate the radiative tail one would normally see in a W distribution, for example.  It is those pesky charge-even and interference pieces that we had to carefully determine.  Not sure, without further study, to what extent the cuts "minimized" these.  We are probably safer to not say anything.
> 
> Page 13:
> - Line 768. Use "Refs. [74,75].".
> Got it.
> - Line 792. Your assigned scale uncertainty of 0.3% (15% of the correction) seems completely
>   arbitrary. Can you justify why this uncertainty for an unknown correlation (as you point out)
>   is reasonable?
> Well, 15% is an  overestimation of the uncertainty on the radiative corrections.  Since this leads to a relatively insignificant systematic uncertainty (compared to the combined total of the other sources) we are "safe" with this number.  I'll try to improve this.
> - Line 844. Use "Charge independence of track reconstruction:"
> Done
> 
> Page 14:
> - Eq.(17) should end with a period for proper punctuation.
> Done
> - Line 887. Spurious "cut" included.
> Fixed
> - Line 888. Need units on z-vertex limits here.
> Fixed
> 
> Page 15:
> - Line 902. Use "Acceptance correction:".
> Done
> - Line 920. Add a period at the end of the sentence.
> Done 
> - My sense of reading your conclusions of the value of the measured TPE effect from this
>   analysis as shown in Fig. 19 is that you are fully biased by what you expect the effect
>   to look like (i.e. by what is needed from TPE effects to explain the form factor discrepancy).
>   I think that this inherent bias affects how you interpret and discuss your results. Regarding
>   Fig. 19 you state (line 939) that the data at Q2=1.45GeV2 shows a moderate epsilon dependence.
>   The data in my opinion do not support this statement by any statistical measure. One data point
>   from your measurement (eps=0.4) is statistically away from unity (and by less than 1 sigma). The
>   other 4 are fully consistent with unity to 3 significant figures.
> I agree that one's eye is biased by the Zhou and Blunden curves.  So perhaps it is better to say "Our data at Q^2=1.45 GeV^2 when combined with the VEPP-3 result show a a moderate eps dependence."  In no way can one expect a constant fit of these results (no eps dependence) to work.  We did say that the Q^2=0.85 show no eps dependence (where our 0.4 point is less than 1 sigma from 1) but at 1.45 GeV our 0.4 data point is nearly 2 sigma away from 1. So with the modification we are not overstating our case.
> - My disagreement with your discussion of the results continues with Fig. 20. Here you state
>   (line 964) that the data are consistent with little or no Q2 dependence. Certainly this does not
>   meet the "eye test" when I look at the data at eps=0.45. I would agree with this qualitative
>   statement only for your eps=0.88 data.
> Remove the curves and the VEPP-3 data and there is no Q^2 dependence to the data. We have 2 data points greater than 1 but just about 1 sigma away from 1. So what exactly is wrong with saying our data show no Q^2 dependence?
> 
> - Fig. 19 caption.
>   - Line 1. Include units after 0.85.
> Done
>   - Line 4. Use "The filled black squares show …".
> Fixed
>   - Next to last line. Use "... figure is a linear fit ...".
> Done
> 
> Page 16:
> - Line 992. Use "Ref. [16]".
> Fine.
> - Line 1008. Use "... we do not include these ...".
> Fixed
> 
> Page 17.
> - Fig. 20 caption.
>   - Line 3. Use "... is our CLAS 2013 result ...". Also include the reference here for the
>     CLAS 2013 result.
> Done
> - Line 1023. Use "Refs. [21,31]".
> Done
> - Table IV caption. Last line. Use "... $\chi^2$ value …".
> Okay
> - Line 1045. Remove comma after "factor".
> Actually, that one should be there and another one after delta.
> 
> Page 18.
> - Ref. [53]. Use "AIP Conf. Proc.".
> That's the way BibTex typeset this.  We need to have something for the copy editors to do.
> 
> Brian
> ********************************************
> Professor Brian A. Raue                    Phone: 305-348-3958
> Graduate Program Director                ​Fax: 305-348-6700
> Department of Physics
> Florida International University
> Modesto Madique Campus
> CP 217
> Miami, FL 33199


***********************************************************************************
*                                                                                                                  *
* Dr. Daniel S. Carman                  e-mail : carman at jlab.org                     *  
* Staff Scientist                              office : (757)-269-5586                         *
* Jefferson Laboratory                   web: http://userweb.jlab.org/~carman  *  
*                                                                                                                  *
***********************************************************************************