The International Standards Organization (ISO) published Technical Specification (TS) 15666 in 2003 to facilitate the comparison of results from surveys on noise annoyance. This document recommends including two standardized questions with two standardized response scales in all surveys of long-term noise annoyance. A recently revised version of the TS proposes a weighting procedure for the response data to achieve a better similarity between the results from the two questions. This paper analyses the results from 43 surveys conducted according to the ISO/TS 15666 recommendations. It is shown that the average difference in the response to the two annoyance questions is equivalent to a 6 dB shift in the exposure level. The proposed weighting of the verbal response whereby response category 5 (extremely) is counted in full, and category 4 (very) is weighted by a factor 0.4 creates a greater similarity between the two responses. It is important to use a definition of prevalence of high annoyance which matches the one used for the development of the relevant noise regulations.