A sample survey question is, "Is smoking on campus an issue?" or "How satisfied are you with our customer service?" Survey questions need to be simple, unbiased and comprehensible in order for respondents to provide valuable, objective answers.
To maintain simplicity, a survey question must only deal with a single concept. Avoid using double-barreled questions in surveys, because questions referring to more than one idea may lead to answers that are open to several interpretations. Ambiguities in answers render researchers incapable of making accurate analyses and drawing conclusions.
Survey questions must remain balanced to avoid producing skewed answers. For example, asking the question, "Do you think abortion is okay if it's the only way for the mother to survive?" is biased. A more balanced equivalent to this question would be, "Do you think abortion is good or bad?" The first question has the tendency to influence respondents to a certain perspective, whereas the second has a more neutral position. Avoid using language with connotations and emotional undertones to avoid bias.
Survey questions must use words that respondents can easily understand. Avoid using technical or academic terms, unless respondents are well-informed in these terminologies. Alternatively, add a thorough description to explain unusual terms before asking a question. Questions that are too technical risk getting either blank or random answers from confused respondents.