10 - H1 - Desy

8.3 Error matrix evaluation 113
evaluated uncertainties. The variables R T , HCF and HCellF are treated simultaneously,
as one may expect that change of the calorimeter shower simulation would
affect all three of them at the same time. The final error due to the propagation
of their uncertainties becomes the leading error due to the fact that small changes
in the width of the cluster may shift events from the Signal unfolding output to
the Background part and vice versa. The error on the total inclusive cross section
has been evaluated as 11%. For a single differential cross section the error varies
between 10% up to 25% in the forward region of the detector. S T , K T and FLF
uncertainties are propagated independently and their influence is much lower (1 -
3%).
• Photon energy scale
The photon energy scale is studied in section 7.2. The alternative migration matrix
is evaluated with 1% correction to the MC photon energy correction for photons
with pseudorapidity η γ < 1.4 and 4% for photons with η γ > 1.4, resulting in the
final error of roughly 1.5% on average.
• Hadronic energy scale
The study of the MC description of hadronic system energy is presented in section
7.3. A 2% correction is applied to the energy of all hadronic particle candidates
in the MC and alternative migration matrix is evaluated. The correction affects
the efficiency loss due to y cut for all cross sections and E jet
T
cut for the exclusive
selection. The error associated with this effect is below 1%.
• The polar angle of the photon
The MC description of the polar angle resolution has been evaluated as 3 mrad [43].
For the pseudorapidity range η γ > 1.4 it is increased to 4 mrad. The propagated
error caused by this uncertainty is included in the final errors, but is of negligible
level.
• Direct to resolved event class ratio
The relative scales of the direct and resolved prompt photon MC production have
been evaluated in section 7.4 together with their uncertainties. Due to lower hadronic
activity for the direct cases one may expect differences in the acceptance corrections,
mostly due to the photon isolation criteria. The calculated error is usually of 1%
order.
• Detector dead material simulation
An uncertainty in the description of the dead material in the simulation is accounted
for by varying the probability of photon conversion before the calorimeter by 10%.
For polar angles θ < 20 ◦ it is varied by 30% because of more dead material in the
forward region of the detector. This results in a 1% error in the central region and
3% in the most forward η γ bin.
All the evaluated systematic error matrices were added and included in the final error
matrix E x . Due to the high sensitivity of the signal extraction to the cluster shape
description, it is by far most significant source of error.