Methods and apparatus to detect spillover in an audience monitoring system are disclosed. An example method includes sampling a first audio signal received by a first microphone. A second audio signal received by a second microphone is sampled. The second microphone is separated from the first microphone by a first distance. A time delay between receipt of the first and second sampled audio signals is calculated. Whether the time delay is within a threshold time difference is determined. When the time delay is within the threshold time difference, the media is identified as local media. When the time delay exceeds the threshold time difference, the media is identified as spillover media.