| 7 |
|
(CSV medium working point as described in Sec.~\ref{sec:selection}). |
| 8 |
|
The sample is dominanted by \wjets\ and is thus used to validate the MC modelling of this background. |
| 9 |
|
|
| 10 |
< |
In Table~\ref{tab:cr1mtsf} we show the amount that we need to scale the Wjets MC |
| 10 |
> |
In Table~\ref{tab:cr1mtsf} we show the amount that we need to scale the \wjets\ MC |
| 11 |
|
by in order to have agreement between data and Monte Carlo in the $M_T$ peak |
| 12 |
< |
region, defined as $60 < M_T < 100$ GeV. These scale factors are not terribly |
| 12 |
> |
region, defined as $50 < M_T < 80$ GeV, for the |
| 13 |
> |
different signal regions. (Recall, the signal regions have different |
| 14 |
> |
\met\ requirements). These scale factors are not terribly |
| 15 |
|
important, but it is reassuring that they are not too different from |
| 16 |
< |
1. [UPDATE WITH TRIGGER EFFICIENCIES] |
| 16 |
> |
1. |
| 17 |
|
|
| 18 |
|
|
| 19 |
|
\begin{table}[!h] |
| 31 |
|
e \mt-SF & $0.94 \pm 0.02$ & $0.90 \pm 0.04$ & $0.84 \pm 0.05$ & $0.80 \pm 0.07$ & $0.83 \pm 0.10$ & $0.77 \pm 0.13$ & $0.86 \pm 0.20$ & $0.87 \pm 0.29$ \\ |
| 32 |
|
\hline |
| 33 |
|
\end{tabular}} |
| 34 |
< |
\caption{ \mt\ peak Data/MC scale factors applied to the single lepton |
| 35 |
< |
samples and \ttdl. The raw MC is used for backgrounds from rare |
| 34 |
> |
\caption{ \mt\ peak Data/MC scale factors applied to \wjets\ |
| 35 |
> |
samples. No scaling is made for backgrounds from other |
| 36 |
|
processes. CR1PRESEL refers to a sample with $\met>50$ GeV. |
| 37 |
|
The uncertainties are statistical only. |
| 38 |
|
\label{tab:cr1mtsf}} |
| 39 |
|
\end{center} |
| 40 |
|
\end{table} |
| 41 |
|
|
| 42 |
< |
|
| 43 |
< |
In Table~\ref{tab:cr1yields} we compare the data and MC yields in the four $M_T$ signal regions |
| 44 |
< |
and in a looser control region. We also derive the data/MC scale factors |
| 45 |
< |
$SFR^{e}_{wjet}$ and $SFR^{\mu}_{wjet}$. The underlying \met\ and $M_T$ distributions |
| 46 |
< |
are shown in Fig.~\ref{fig:cr1met} and~\ref{fig:cr1mtrest} |
| 47 |
< |
|
| 42 |
> |
Next, in Fig~\ref{fig:cr1met},~\ref{fig:cr1mtrest}, |
| 43 |
> |
and~\ref{fig:cr1mtrest2}, we show plots of \met\ and then $M_T$ |
| 44 |
> |
for different \met\ requirements corresponding to those defining our signal regions. |
| 45 |
> |
It is clear that there are more events in the $M_T$ tail than |
| 46 |
> |
predicted |
| 47 |
> |
from MC. This implies that we need to rescale the MC \wjets\ |
| 48 |
> |
background |
| 49 |
> |
in the tail region. |
| 50 |
|
|
| 51 |
|
\begin{table}[!h] |
| 52 |
|
\begin{center} |
| 95 |
|
\hline |
| 96 |
|
\hline |
| 97 |
|
\hline |
| 98 |
< |
$SFR_{wjet}$ & $1.48 \pm 0.11$ & $1.64 \pm 0.20$ & $1.38 \pm 0.24$ & $1.26 \pm 0.36$ & $0.96 \pm 0.45$ & $1.02 \pm 0.67$ & $1.23 \pm 0.91$ & $1.12 \pm 1.31$ \\ |
| 98 |
> |
$SFR_{wjet}$ & $1.48 \pm 0.26$ & $1.64 \pm 0.38$ & $1.38 \pm 0.30$ & $1.26 \pm 0.39$ & $0.96 \pm 0.45$ & $1.02 \pm 0.67$ & $1.23 \pm 0.92$ & $1.12 \pm 1.31$ \\ |
| 99 |
|
\hline |
| 100 |
|
\end{tabular}} |
| 101 |
|
\caption{ Yields in \mt\ tail comparing the MC prediction (after |
| 102 |
|
applying SFs) to data. CR1PRESEL refers to a sample with $\met>50$ |
| 103 |
< |
GeV and $\mt>150$ GeV. |
| 104 |
< |
The uncertainties are statistical only. |
| 103 |
> |
GeV and $\mt>150$ GeV. See text for details. |
| 104 |
> |
% The uncertainties are statistical only. |
| 105 |
|
\label{tab:cr1yields}} |
| 106 |
|
\end{center} |
| 107 |
|
\end{table} |
| 108 |
|
|
| 109 |
|
|
| 110 |
+ |
The rescaling is explored |
| 111 |
+ |
in Table~\ref{tab:cr1yields}, |
| 112 |
+ |
where we compare the data and MC yields in the $M_T$ signal regions |
| 113 |
+ |
and in a looser control region. Note that the |
| 114 |
+ |
MC is normalized in the $M_T$ peak region by rescaling |
| 115 |
+ |
the \wjets\ component according to Table~\ref{tab:cr1mtsf}. |
| 116 |
+ |
|
| 117 |
+ |
We also derive data/MC scale factors. |
| 118 |
+ |
As shown in Table~\ref{tab:cr1yields}, these are derived in two different ways, separately for muons and |
| 119 |
+ |
electrons and then combined, as follows: |
| 120 |
+ |
\begin{itemize} |
| 121 |
+ |
\item For the first three sets of scale factors, above the triple horizontal |
| 122 |
+ |
line, we calculate the scale factor as the amount by which we would |
| 123 |
+ |
need to rescale {\bf all} MC (\wjets\ , \ttbar\ , single top, rare) in |
| 124 |
+ |
order to have data-MC agreement in the $M_T$ tail. |
| 125 |
+ |
\item For the next three set of scale factors, below the triple horizontal |
| 126 |
+ |
line, we calculate the scale factor as the amount by which we would |
| 127 |
+ |
need |
| 128 |
+ |
to scale \wjets\ keeping all other |
| 129 |
+ |
components fixed in order to have data-MC agreement in the tail. |
| 130 |
+ |
\end{itemize} |
| 131 |
+ |
\noindent The true \wjets\ scale factor is somewhere in between these |
| 132 |
+ |
two extremes. We also note that there is no statistically significant |
| 133 |
+ |
difference between the electron and muon samples. We use these data |
| 134 |
+ |
to extract a data/MC scale factor for \wjets\ which will be used to |
| 135 |
+ |
rescale the \wjets\ MC tail. This scale factor is listed in the last |
| 136 |
+ |
line of the Table, and is called $SFR_{wjets}$. It is calculated as |
| 137 |
+ |
follows. |
| 138 |
+ |
\begin{itemize} |
| 139 |
+ |
\item Separately for each signal region |
| 140 |
+ |
\item As the average of the two methods described above |
| 141 |
+ |
\item Including the statistical uncertainty |
| 142 |
+ |
\item Adding in quadrature to the uncertainty one-half of the |
| 143 |
+ |
deviation from 1.0 |
| 144 |
+ |
\end{itemize} |
| 145 |
+ |
|
| 146 |
+ |
|
| 147 |
+ |
|
| 148 |
+ |
|
| 149 |
+ |
|
| 150 |
+ |
|
| 151 |
|
\begin{figure}[hbt] |
| 152 |
|
\begin{center} |
| 153 |
|
\includegraphics[width=0.5\linewidth]{plots/CR1plots/met_met50_leadmuo_nj4.pdf}% |
| 201 |
|
\end{figure} |
| 202 |
|
|
| 203 |
|
|
| 204 |
< |
\clearpage |
| 204 |
> |
\clearpage |
