Within the realm of single response survey question lays the domain of the mystics. Okay perhaps it is not the Lord of the Rings, but numerical questions, including the famous Likert Scale, can generate very actionable data if constructed properly.
The respondent is asked to select a single response from a scale, typically 1 – 5, 1 – 7, 1 – 9, or 0 – 10. The power of this type of question comes when several are linked together to provide a more holistic viewpoint of the attitude under consideration. Multiple-item measures have higher levels of validity of compared to single-item measures. For example, customer satisfaction of an ocean cruise vacation could be measured by a single item (overall satisfaction), but it becomes more actionable when the overall measure is combined with items examining the various aspects of the cruise experience.
The graphic below came from a recent satisfaction survey. The vendor in question was the auto dealer who serviced my car. The instruction it provides prepares the participant to respond to the numerical questions which follow. It reinforces the notion that it is important to rate the aspects individually and that variation in responses between items is acceptable. This will help to minimize errors associated with straight-lining (selecting responses all from one side of the scale or the other).
The second graphic shows the three items along with the scale points. Using both numerical and verbal anchors has merit. This approach appeals to both numerical and verbal thinkers. Again the instructions at the top of the items remind respondents to think of each item independently.
In this graphic the topic of interest is service initiation. Thinking from a multi-item perspective consumer and B2B market researchers should strive to incorporate all of the relevant dimensions of the consumer experience. This may require the use of exploratory qualitative research to identify the dimensions that are critical for satisfaction, re-purchase, or other relevant research concerns.
Single or multiple-item measures allow the researcher to access a broader range of statistical analysis techniques, thus increasing the power of the data. They require more cognitive resources on the part of the respondent than binary response measures (yes/no or selected/not selected), but if used properly and in moderation they can up your research game.
Photo Credit: Ambro