Cochlear reflectance (CR), the cochlear contribution to ear-canal reflectance (ECR), has theoretical advantages for cochlear modeling. Comparisons between measurements and models may lead to improved interpretation of cochlear status and provide a basis for making improvements to the models. Previous evaluation of clinical utility of CR measurements showed that CR did not predict auditory status or behavioral threshold as accurately as other otoacoustic emission measurements. Strategies for improving the quality of CR measurements were assessed in ECR measurements from 27 participants. Results indicate that the quality of CR measurements can be improved by (1) increased averaging time and (2) adjustment to the methods for extracting CR from ECR. Simulation of ECR was performed using a combination of a middle-ear model and a one-dimensional cochlear model. Simulated CR was the ECR difference between active and passive conditions of the model. The model simulation results were compared with measurements of ECR and CR in both the time and frequency domains. Minor disparities between measurements and model will provide a basis for improvements in the model. Substantial agreement between measurements and model suggest that CR is consistent with linear coherent reflection due to random impedance perturbations along the cochlear partition.