The invention relates to methods and apparatus for performing touch-screen surveys. The surveys comprise a plurality of screen displays each including a survey question and one or more screen buttons by means of which the user signifies an answer to the survey question by touching the screen button. A method of quarantining responses to a touch screen survey comprises defining a set of criteria for identifying if a response is likely to be a serious response or a nuisance response. The method includes monitoring the screen touches to determine the number of survey questions answered and the time intervals between screen touches. The monitored screen touches are compared with the defined criteria to categorize the response as a likely nuisance response or a likely serious response. The response is quarantined if it is categorized as a likely nuisance response. A method of discriminating between valid and nuisance responses comprises: monitoring screen touches; determining a same place touch count depending on a zone at which the screen is touched and the number of consecutive touches in the zone; comparing the same place touch count to a threshold; and determining if the response is valid or a nuisance response based on the comparison.