Survey Design

After identifying the problem to be addressed, the next step is to design a survey that targets the correct respondents, asks appropriate questions, and minimizes biases. There are several industry-standard best practices to keep in mind during this stage, as well as a few other factors to consider.

About the links in “Survey Design”

  • Mutually Exclusive Collectively Exhaustive

    Mutually Exclusive, Collectively Exhaustive (MECE) refers to the concept of organizing information such that it is separated into buckets that are both mutually exclusive.

  • Survey Logic

    Survey Logic controls how questions within a survey relate to one another, allowing a respondent’s answer to a question to affect other questions they will or will not be shown.

  • Quotas

    Most research requires responses from certain groups, rather than an entirely random sampling of the population. In this situation, studies use quotas to cap certain categories of respondents.

  • Census Representative

    Used to ensure that the distribution of respondents accurately reflects the population of the country.

  • Randomization

    Randomization is used to remove or reduce serial position biases that can result in over or underrepresentation of certain answers based on their position within a list.

  • Likert Scale

    A Likert scale is a widely-used rating system for survey research, consisting of a symmetrical scale of positive and negative responses.

  • Monadic Testing

    Monadic Testing is an experimental design that allows researchers to create a single testing instrument that systematically assesses multiple different concepts and controls for the potential for bias.

  • Semantic Differential Scale

    Semantic Differential Scales are used to quantitatively measure attitudes and perceptions regarding people, brands, activities, etc. A typical semantic differential scale is a 7-point scale, with a neutral point in the middle.