Is the poll based on a random sample of the entire population?
A truly random sample requires that everyone in the entire target population (usually all adults) have an equal chance of being interviewed. Typically, a “good” random sample size is somewhere between 1,000 and 1,500 persons who are selected in a “stratified” or multi-stage process where progressively smaller geographical units are randomly selected as sample areas. Most polls will describe their sample methodology in the fine print of their results. Those where the pollsters seek out respondents are likely to be scientific, while those that allow respondents to select themselves are usually not. Examples of unscientific polls include call-in polls, Internet polls, and mall surveys. • Are the questions understandable? The questions must be framed in basic language that people can understand. This means that pollsters avoid terms and jargon that require further explanation. If some explanation is needed, the pollsters should present it in a basic and balanced way. In addition, questions
