Assessment of human exposure to RF signals which are rapidly variable in time and wide-banded is not yet standardized. Signals in IEEE 802.11 standards are widespread among population use due to presence of wireless local area network in everyday life. Recently, with technological advancement of real-time vector signal analyzers, a more precise exposure evaluation became achievable. However, such instruments are prohibitive and not available for the majority of metrological laboratories. Therefore, it is of great value to improve accuracy obtained by using swept-spectrum analyzers through procedural adequacy. This paper contributes to the refinement of the measurement procedure of stochastic-like radio signals level by a solution of averaging which responds to the dual variability character of such signals: in time and in amplitude.