[Mott] [EXTERNAL] Tim recants

[Charles Sinclair]

Dear Tim,

I confess to being a little surprised that you only now discovered that 
people often add systematic and statistical uncertainties in 
quadruture.  I don't know where you learned in your youth, but I thought 
you went to grad school at Yale, working with Vernon Hughes - a noted 
scientist.  I, too, worked with Vernon, and can personally attest (and 
show co-authored papers) with statistical and systematic uncertainties 
combined in quadrature.  I share with NIST a lack of enthusiasm for 
terms like statistical and systematic "errors", preferring the more 
accurate (to my mind) use of "uncertainties". As for the issue of 
combining these very different origins of such uncertainties, one might 
look to what is now common in much of high energy physics presentations 
these days.  Here, one simply states the experimentalist's values for 
the statistical and systematic uncertainties.  With high energy physics' 
reliance on incredibly sophisticated codes for experimental 
interpretation these days, this is arguably best.  I note that in the 
push to exceptionally high precision in some measurements (e.g. NIST 
type stuff), it is also becoming common to simply state statistical and 
systematic uncertainties separately and leave it at that.  Perhaps we 
should do this for Mott?  I would be OK with that, and I suspect that if 
we ever do another "spin dance" at JLab, the choice may well be to 
simply state the statistical and systematic uncertainties separately and 
leave it at that.  Not overly satisfying to the reader, to be 
sure..............

Best,
Charlie

On 2/9/2020 1:00 PM, Timothy Gay wrote:
>
> Dear Mott team:
>
> I may have spoken hastily on my last point emailed last night.  I just 
> dug in to this more thoroughly, and the NIST Website on Uncertainty
>
> https://urldefense.proofpoint.com/v2/url?u=https-3A__physics.nist.gov_cuu_Uncertainty_international1.html&d=DwID-g&c=CJqEzB1piLOyyvZjb8YUQw&r=hIi4A-HgNXaCb56eFCTIew&m=WNwXjqkuXRm6edFa6GrGbg6mPQLLGLj13biWtQCKPCg&s=gzBj8O-XzMiHMta2NKNU0Se0my0JLq2cpJi8AJRYTJQ&e=  
> <https://urldefense.proofpoint.com/v2/url?u=https-3A__physics.nist.gov_cuu_Uncertainty_international1.html&d=DwMFAg&c=CJqEzB1piLOyyvZjb8YUQw&r=hIi4A-HgNXaCb56eFCTIew&m=zMBOY7vHu-mMwNo1MpaVDXoP9hrfczgc7quQ164i00U&s=p2uPrKrKpeXld665zEELAMpibURJhXBAJoAlXljn4U0&e=>
>
> seems to imply that adding systematic and random uncertainties in 
> quadrature is OK (although they discourage the use of the phrase 
> “systematic error”).  So maybe Table 3 is OK as it stands.  (This is 
> not what I learned in my youth!)
>
> Tim
>
> *From:*Mott [mailto:mott-bounces at jlab.org] *On Behalf Of *Joe Grames
> *Sent:* Thursday, February 6, 2020 2:20 PM
> *To:* mott at jlab.org
> *Subject:* [Mott] Version #8
>
> Hi Mott Team,
>
> Please find attached version #8 of the Mott paper.  The attached copy 
> address ~all of the edits we agreed to during our conference call a 
> month ago, thanks especially to Marcy and Daniel for polishing off the 
> fitting section.   A short while after that meeting, Charlie asked 
> that we take one last hard look at how we've categorized and tallied 
> the uncertainties.  Riad and Daniel largely championed this 
> assessment, and consequently the paper is better for it.   While there 
> is no new surprise twist ending, the uncertainties and studies therein 
> of what we did are clearly and correctly spelled out.
>
> If you would like to make some mild polishing, fix grammar or 
> punctuation that's okay, just edit in review mode and return to me by 
> *Monday, Feb 10*.  Note, I'm not asking for any further edits, and 
> would not fret over layout, fonts, etc, b/c that is dictated by the 
> publisher.  However, if you find something objectionable then probably 
> a good idea to reply-all and make your case.  Short of anything that 
> slows this bus down, I would like to submit the paper soon after Monday.
>
> Joe
>
>
>
> _______________________________________________
> Mott mailing list
> Mott at jlab.org
> https://mailman.jlab.org/mailman/listinfo/mott

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.jlab.org/pipermail/mott/attachments/20200209/d1e9c88d/attachment-0001.html>