Two versions of a question on impact sounds from neighbouring spaces in dwellings have been demonstrated to return different results. At first, a new version of a questionnaire appeared to be imprecise, too broad etcetera and therefore it returned higher annoyance ratings than the previous, more specific question. But the previous version may have been too restricted - disturbing sounds other than footstep sounds in building with concrete floors seem be a greater problem than anticipated. If so, concrete floors may have appeared to perform better in the previous surveys than occupants actually experience. A modified questionnaire is used in a current survey, which asks for specific sources of impact sounds.