A market research survey can be an effective way to find out what customers are looking for or how they might respond to a new product, but the effectiveness of the survey depends on the accuracy of the results. Psychologists who use surveys in their research have found that the wording of the questions can affect how people respond.


If the wording of a survey suggests that one answer would be more socially acceptable than another, some respondents may answer inaccurately. For example, if a craft brewer conducts a survey about a proposed new beer product and asks "is flavor or alcohol content more important to you," some respondents might be embarrassed to say that alcohol content is more important to them even if it is. This type of response bias is called socially desirable responding. Market researchers can minimize the effects of SDR by phrasing all questions in a neutral way. The effects of SDR are stronger for in-person interviews.


Another type of response bias is yea-saying or acquiescence. Some surveys are structured as a series of statements with which the respondent is asked to agree or disagree, such as "traditional thin-crust pizza is best." Some respondents tend to agree to every statement without really thinking about any of them. Market researchers try to limit the effects of yea-saying by including statements that contradict each other. The survey about pizza crust could include the statement "Chicago-style deep-dish pizza is the most appealing." If the same person agrees with both statements, this is evidence of acquiescence and the answers should be considered unreliable.

Extreme and Midpoint Response

If a survey includes statements such as "please rate thin-crust pizza on a scale of one to five where one means that you don't like it at all and five means you like it very much," some respondents will answer either one or five to all such questions, avoiding any moderate responses. This is called the extreme response bias. Some respondents take the opposite approach and answer every question of this type with a rating of three. This is called the midpoint response bias. Some market researchers try to get around both biases by asking multiple-choice questions or questions with only two options such as like and dislike.


Adjectives can also affect the way people respond. For instance, a harsh phrasing such as "do you enjoy heavy and greasy deep-dish pizza" may encourage respondents who would actually enjoy that type of pizza to claim that they would not. A phrasing such as "would you enjoy a new gourmet pizza featuring delicious Asiago cheese" might encourage some respondents to say that they would be interested in the product when they have actually never tasted Asiago cheese and don't know whether they would like it or not.