The research examined the reliability of environmental noise (EN) data by varying the measurement period of EN from 1, 3, 5, to 7 days. EN data were collected at four sites in urban and rural areas at 10-minute intervals for more than one year. Using these data, the mean and standard deviation of the EN for different measurement periods were calculated at each site. The probability that the EN would be erroneously judged to have "not increased in noise level" was calculated from the EN determination if a steady noise or a 3 dB louder noise with a 1 dB variation was added to the EN. As a result, it clarifies that in locations where the standard deviation is large, not only should the number of days of measurement be increased, but also measures should be taken to avoid misjudgments depending on the measurement timing.