There are many problems with drawing inferences about overall public opinion from programmes like BBC Question Time. The main one is sample bias. Sampling occur

There are many problems with drawing inferences about overall public opinion from programmes like BBC Question Time. The main one is sample bias. Sampling occurs in three stages: 1) Audience members are self-selecting. 2) Audience members are then selected from the group that self-selected to represent a cross-section of the population. 3) Not all audience members get to ask their question, they are selected by the host. The problem with the second stage is that there is a person, called an "audience producer", who selects people. When that person reposts the social media output of Britain First, an extremely far right political organisation, it questions whether they may have a political bias that may affect the sampling [0]. Watching Question Time could quite easily convince you that you are in a majority or a minority and you could be wrong. Treating Question Time as any kind of meaningful measure of public opinion is a form of self delusion, albeit one that's hard to shake off even if you know it. The only thing a person speaking on Question Time could tell you is that there exists a person with those views, and even that might not be true. Same goes for any other audience panel. [0] https://www.opendemocracy.net/ourbeeb/tim-holmes/is-question-time-s-audience-producer-really-fascist https://www.opendemocracy.net/ourbeeb/tim-holmes/is-question...