By Jeffrey Henning
Act 1 – Setting: A Marketing Researcher’s Office
Charlie: How’d that last survey go?
Jessie: Well, the results were pretty clear that consumers preferred Concept C, but it was so different from what the brand had done before that they went with Concept A instead. The brand managers trusted their gut and their past experience more than the research.
Charlie: That’s too bad. Happens sometimes. Well, let’s see about this new questionnaire you submitted.
Jessie: It’s for a new concept test.
Charlie: OK, here you’re using a numeric scale with just the endpoints labeled. According to Krosnick and Tahk in “The Optimal Length of Rating Scales to Maximize Reliability and Validity”, such scales are less consistent, especially with lower educated participants. Let’s use a five-point fully labeled scale, instead. It’s more reliable, according to the research.
Jessie: Well, I always use the 0 to 10 point scale. I’m just more comfortable with it. So let’s keep it.
Charlie: But the research shows…
Jessie: It’s always worked for us in the past. No need to change.
I have a conversation like the above at least once a week. Researchers aren’t so different from the users of our research. It’s easy to ignore research on research in favor of past practices. That’s why I teamed up with QuestionPro to create this video walking people through best practices for interval rating scales. Enjoy!