From stuart@neutrino.physics.ucsb.edu Wed Jul 16 18:32:18 2003
Date: Fri, 11 Jul 2003 14:52:48 -0700 (PDT)
From: David Stuart 
To: smaria@fnal.gov, burkett@fnal.gov
Cc: williams@fnal.gov, castro@fnal.gov, slaughter@fnal.gov,
     affolder@neutrino.physics.ucsb.edu, chill@fnal.gov, incandel@fnal.gov
Subject: Comments on first draft of Kaluza-Klein Graviton Emission paper


Dear authors and godparents,

The UCSB group has read the draft jets+met PRL. We congratulate you on
bringing this analysis to the publication stage. Our comments on the
draft and questions on the analysis are listed below.

- David, for the UCSB group

Abstract: The date of the running period should be included in the
abstract to clarify that this is a Run 1 result.


Response
We believe that saying sqrt(s)=1.8 should make it clear and the running period is mentioned in the main text.

Introduction: The theoretical introduction is quite long. While it is
informative and well-written, the space could be better used for an
expanded description of the experiment itself. For example, the figure
with the feynman diagrams could be removed.

Response
We have expanded the experimental part to address all the comments.
Since this is a dedicated KK search we think it is OK to keep the graphs.


Page 4: The explanation of why 2 jet events are used should be expanded.
It might be easiest to explain it after the discussion of the backgrounds.

Response
OK.  We have moved it and changed the wording a bit.  We have already
had some iterations with the GPs on this.

Page 4: The way in which W and QCD backgrounds mimic the signal should be
explained. It may be obvious to us but not to a typical PRL reader.

Response
OK.  We have added this explanation.

Page 5: From the supporting note, we see that you used CTEQ2. That is a
rather old PDF with known deficiencies. We don't think it is necessary to
repeat a study with a recent PDF, but the PRL should note that CTEQ2 was
used and the systematic uncertainty assigned to PDF choice.
Response
CTEQ2 is used only for the cross checks.  The default PDF is
MRSD-.  We have added a reference that this is what was used.


Table II: Please show the number of events observed in this table. It is
key to the measurement.

Response
OK.  Fixed.


Fig 2 and 3: We think that the expected signal shape should be included in
these figures. The plots show that the data matches the expected
background. The reader should also be shown how it would reveal a signal
if it were present.

Response
It would be a smooth excess over the backgrounds for most of
the kinematic distributions.  In CDF5151 we have plots (also on our webpage
at (http://harv11.fnal.gov/~burkett/KKG) that show the addition of the signal
for M=1TeV and n=2.  
The plots with the signal added become too busy to be easily read.  In talks 
we have emphasized the difficulty of trying to observe the signal in the MET 
distirubtion.  In the latest draft we have added a sentence to explain this.


Reference 12: d_0^i must be a typo.
Response
OK.  Fixed.

Reference 16: The D0 paper now has a PRL reference which should be used
instead of the hep-ex reference.

Response
 OK.  Fixed.


We also have the following questions/comments about the analysis which we
did not find (or missed) in your note.

- The shape and normalization of the irreducible Z->nunu + jets background
is key to this search. (It is an impressive feat that irreducible physics
backgrounds dominate!) You use the Z->ee + jets data to normalize the
monte carlo, but there is a significant discrepancy at low ET. To what
extent is that important for this result? Could Z-> mumu help as a
cross-check?

Response
  The discrepancy in Z_PT is actually (missing) physics in the 
simulation (sudakov) together with resolution.  This much could be checked 
with muons, since tracking is better at low pt. Also look at p.47 of cdf4512.  
Note that the discrepancy is much smaller when you actually compare the met 
spectra.  We checked the normalization after throwing away the first two 
bins and the result differs by less than 5%. The actual slope of the falling 
missing et spectrum matches quite well which is important. 

For the Zmumu: No-- remember that in run Ib we did not even attempt 
the ratio of W/Z in muons because the trigger was screwed up 
altogther and the numbers made no sense (ratios of 14-20 etc). 


- For the QCD background, b -> l nu is not well modeled in herwig and
could contribute real MET. How is the contribution of real MET (i.e., not
from jet mis-measurement) in QCD backgrounds determined?

Response
bbar and ccbar are less than 10% of the QCD estimates, and the final
QCD background component is pretty small.  Also, we do have cleo/qq in 
the MC with the closest to accurate b decays.