We present Monte Carlo modeling studies to provide a quantitative understanding of contrast observed in spatially resolved reflectance spectra of normal and highly dysplastic cervical tissue. Simulations have been carried out to analyze the sensitivity of spectral measurements to a range of changes in epithelial and stromal optical properties that are reported to occur as dysplasia develops and to predict reflectance spectra of normal and highly dysplastic tissue at six different source-detector separations. Simulation results provide important insights into specific contributions of different optical parameters to the overall spectral response. Predictions from simulations agree well with in vivo measurements from cervical tissue and successfully describe spectral differences observed in reflectance measurements from normal and precancerous tissue sites. Penetration depth statistics of photons detected at the six source-detector separations are also presented to reveal the sampling depth profile of the fiber-optic probe geometry simulated. The modeling studies presented provide a framework to meaningfully interpret optical signals obtained from epithelial tissues and to optimize design of optical sensors for in vivo reflectance measurements for precancer detection. Results from this study can facilitate development of analytical photon propagation models that enable inverse estimation of diagnostically relevant optical parameters from in vivo reflectance measurements.