Move Responders to the Top-Box
What Is a Top-Box Survey Response?
By convention, survey responses are classified as top-box, bottom-box, and middle-box scores. The top-box is the most positive response to a survey question, and the second-most top-box is the next highest (most positive) response. If respondents are asked to use a Likert scale of "1" through "5" with "5"being the most positive response, then ratings of "4" and "5" are referred to as the two top-boxes.
Typically, a Likert scale is created with five or seven rating "points." The high end of the scale -- responses marked "5" -- will represent words like "very," "highly," "most", and "best." Responses marked "4" are also considered on the high end of the scale and will be associated with words like "slightly" and "better." The middle range of the scale where responses are marked "3" is intended to convey a neutral or non-committal response. The low end of the scale where responses are marked "2" is also associated with words like "slightly" or "very," but the scale points will include a second less-positive word. For example, a rating of "2" could stand for a response indicating that a survey participant is "slightly dissatisfied" or "somewhat disagrees." On the lowest end of the scale, a rating of "1" indicates that the survey participant is "very dissatisfied" or "strongly disagrees."
The box ratings are helpful for analyzing and summarizing data. For instance, suppose the summarized scores on a survey show that 30% of the ratings are in the top-box, which is equivalent to a Likert scale rating of "5" or very highly satisfied, and 15% of the ratings are in the second-most top-box, which is equivalent to a Likert scale rating of "4" or highly satisfied.
The market researcher could report that the survey respondents were highly satisfied because a Likert scale rating of "3" would be the equivalent of satisfied, and 45% of the respondents are more than just satisfied.
A complete list of 12 techniques for creating dynamic survey questions follows. Immediately below is a quick tip for creating delayed branching questions in your surveys research.
Delayed Branching in the Construction of Dynamic Surveys Research
The technique known as delayed branching is similar to compound branching as multiple criteria can be used to configure the branching. The important--and evident--difference is that delayed branching does not occur immediately following the collection of the data upon which the configuration is constructed. Delayed branching can occur anywhere in the the survey as long as the data that establishes the criteria has already been collected.
A common use for delayed branching is when a text response is desirable, but the market researcher does not want to disrupt the flow of several linked or related questions. Additionally, survey respondents will sometimes terminate the survey if they are asked to provide lengthy text responses, or if they perceive that there are too many items in a survey. A market researcher might use delayed branching to increase the probability that the respondent has explored the more important items and answered those questions first.
Although survey questions designed to follow the 12 configurations below are most often seen in online surveys, the concepts can also be applied to paper-based or oral surveys.