This preview shows page 1. Sign up to view the full content.
Unformatted text preview: program. Both D0
(Abazov et al. (2007e) and CDF (Aaltonen et al., 2008f)
published results in this ﬁnal state. The D0 publication uses a
data sample corresponding to R dt ¼ 0:45 fbÀ1 . The CDF
published results correspond to Ldt ¼ 1:0 fbÀ1 .
The initial selections for both experiments require events
with two leptons whose invariant mass is consistent with a Z
boson and two additional jets. The backgrounds in this
channel arise from Z þ jets, tt events, diboson production,
and from jets misidentiﬁed as leptons or a lepton þ jet system
being misidentiﬁed as an isolated lepton. They are determined using a combination of simulated events and data
control samples in essentially the same manner as for the
WH channel. One difference with respect to the WH
Rev. Mod. Phys., Vol. 84, No. 4, October–December 2012 calculation is that sidebands of the ‘‘ invariant mass distribution are used by D0 to determine the misidentiﬁcation and
false isolation background. Both experiments identify b jets
using secondary vertex algorithms.5 The CDF result splits the
ﬁnal tagged sample into single- and double-tagged samples as
in the WH analysis, but the D0 result does not. The D0 result
is derived using the dijet mass distribution only, while the
CDF result is based on the binned likelihood of a NN output.
The sources of systematic uncertainties for these channels are
the same as for the WH result. Table XXXIII shows the yields
for the single- and double-tag analyses from CDF, and
Table XXXIV shows the yields for the dijet and doubletagged samples for D0, respectively. Figure 54 (Fig. 55)
shows the dijet invariant mass (neural network output) distribution from the D0 (CDF) analysis. This distribution is used
as the input to the limit setting program.
Systematic uncertainties are included for trigger and lepton
identiﬁcation efﬁciencies, parton density functions, background cross sections, b-tagging efﬁciencies, jet energy reconstruction, and the methods used to estimate instrumental
backgrounds. The CDF result also includes a systematic from
the top mass uncertainty. The systematic uncertainties for the
D0 measurement range between 2% and 20% expressed as a
fraction of the total background. The systematic uncertainties
for the CDF measurement range between 1% and roughly
25% of the total background. The largest contribution for
both experiments is from the background cross sections. 5 For D0 this differs from the WH case because the published ZH
results predate the availability of the NN tagger. More recent D0
preliminary results use the same NN tagger described in the WH
analysis. Hobbs, Neubauer, and Willenbrock: Tests of the standard electroweak model at . . . 1517 10 2 σ(95% CL) (pb) Expected Limit
Observed Limit 10 1 10 -1
100 FIG. 54 (color online). Dijet mass distribution for the D0 1:0 fbÀ1
ZH ! ‘‘bb analysis. The ﬁnal limits, derived using the same methods as the
WH results, are shown for CDF in Fig. 56 and for D0 in
Fig. 57. As...
View Full Document
This document was uploaded on 09/28/2013.
- Fall '13