MINOS is a long-baseline neutrino oscillation experiment. It consists of two large steel-scintillator tracking calorimeters. The near detector is situated at Fermilab, close to the production point of the NuMI muon-neutrino beam. The far detector is 735 km away, 716 m underground in the Soudan mine, Northern Minnesota. The primary purpose of the MINOS experiment is to make precise measurements of the "atmospheric" neutrino oscillation parameters (Δm²_atm and sin²2θ_atm). The oscillation signal consists of an energy-dependent deficit of v_μ interactions in the far detector. The near detector is used to characterize the properties of the beam before oscillations develop. The two-detector design allows many potential sources of systematic error in the far detector to be mitigated by the near detector observations. This thesis describes the details of the v_μ-disappearance analysis, and presents a new technique to estimate the hadronic energy of neutrino interactions. This estimator achieves a significant improvement in the energy resolution of the neutrino spectrum, and in the sensitivity of the neutrino oscillation fit. The systematic uncertainty on the hadronic energy scale was re-evaluated and found to be comparable to that of the energy estimator previously in use. The best-fit oscillation parameters of the v_μ-disappearance analysis, incorporating this new estimator were: Δm² = 2.32^+0.12_-0.08 x 10^-3eV² sin²2θ > 0.90 (90% C.L.). A similar analysis, using data from a period of running where the NuMI beam was operated in a configuration producing a predominantly V‾_μ beam, yielded somewhat different best-fit parameters Δ‾m² = (3.36^+0.46_-0.40 (stat.}) ± 0.06(syst.)) x 10^-3eV², sin²2‾θ =0.86^+0.11_-0.12(stat.) ± 0.01(syst.). The tension between these results is intriguing, and additional antineutrino data is currently being taken in order to further investigate this apparent discrepancy.