Data-driven Background Estimation for the One-lepton SUSY Search Mode in ATLAS.

In this thesis a search strategy for Supersymmetry within $\int\mathscr{L}dt$ = 1 fb$^{-1}$ of ATLAS data is described. The focus is on inclusive signatures with multiple high-p$_T$ jets, missing transverse energy and one isolated lepton in the final states. The discovery potential is evaluated for several benchmark points in minimal Supergravity parameter space using Monte-Carlo simulations with next-to-leading order cross-section corrections. Further, studies of data-driven methods for the estimation of the Standard Model backgrounds in this particular search channel are carried out. The performance of a two-dimensional sideband-based approach is evaluated and found to be dominated by systematic effects. A new, analytical data-driven background estimation technique ('4-tiles'-method) is introduced and tested extensively using toy Monte-Carlo simulations. Stable and accurate results are obtained. The discovery potential for all studied benchmark points using this method is re-evaluated, including the treatment of systematic uncertainties.