Share what you know with millions of people
Focus is the best place to turn what you know into remarkable content
0
Survey methodology: what is your preferred number of points on a rating-scale?
Researchers have many theories surrounding the 'proper' number of points on a rating scale, often depending on the type of research conducted. For instance, some claim 5 point scales are inappropriate specifically for customer satisfaction surveys because the nuance between 'somewhat' and 'very' satisfied is easily diluted, relative to a 7 or 10 point scale. My question is: as a respondent, what do you prefer? 3, 5, 7, 9, 10, 11 point? As a researcher, what do you think delivers the best value? When? Why?
Events
- Dos and Don'ts of Small Business Marketing May 29 @ 11 am PT
- Lead Nurturing 202: The Next Generation May 31 @ 11 am PT
- The Tricks to Paid Media June 6 @ 11 am PT
- Display Advertising for Brand Awareness June 20 @ 11 am PT
Related Activity
6 Answers
You're referring to a Likert Scale, and normally we see somewhere between 4 and 10 point scales. Some psychologists disagree that the 5-point scale is the most effective, but it's one we see frequently, and it's the one I prefer. I use it the most in developing my surveys, and I like using it when I am taking a survey. So, to your question, I strongly agree that the 5-point scale is my preference from both perspectives.
If you want to see how much this issue gets bandied about, just check Wikipedia: http://en.wikipedia.org/wiki/Likert_scale
I agree with Eric. I too prefer the 5 point scale. If 3 is average on the items tested then you have two levels up (4 and 5) and two down (2 and 1). It really forces those who are a little "wishy-washy" to really pin their answers down. The problem with the ten point scale is that its off from the very beginning.
If 5 is average then you have four lower (4, 3, 2, 1) but you have FIVE higher (6, 7, 8, 9, and 10). This weights the scale in favor of positives and lessens the negatives. You could use nine or eleven but you would have the same problem, the mid point in 11 would be five and a half, and the consumer would be confused because they are used to "Top Five", "Top Ten" etc.
You have to trade off sensitivity (lots of points) against accuracy (how often do respondents pick 7 when they "really" mean 8). Lots of research in this area suggests that, for most surveys, somewhere between 5 and 10 points is ideal by this criterion.
Customer satisfaction data is almost always skewed towards the top end. If you couple that with the tendency that some people have to avoid the top box, then you can get a wide range of feelings all squeezed into one box on a five point scale - the "courtesy 4". Using a 10 point scale allows for more differentiation at the top end of the scale, where the action is.
You also need to think about whether to use a fully verbal scale or a numerical scale, with just the endpoints anchored with words. Verbal scales with more than 5 points are pretty cumbersome to use, especially over the telephone.
10 point numerical scales are easy for respondents - giving a score out of 10 is a natural thing to do. They also have beneficial properties when it comes to analyse them. For those reasons, I'd vote 10 point numerical for customer satisfaction surveys.
One final thought - in event-driven or transactional surveys it's often preferable to use "Yes/No" questions where possible instead of rating scales, with a 10 point overall satisfaction question as an outcome variable.
I am a big fan of a 5 point scale. You are typically seeking subjective feedback. Customers not only have a challenging time differentiating between a 7 and an 8 on a 10 point scale, in the end, the difference probably does not matter. Additionally, it is best if the semantic differentials sound like real people talk, not like a researcher wrote. Also, keep in mind that evaluating something subjective like an experience deserves emotional language. Otherwise you get: Question: "Overall, how would you rate your honeymoon experience?" Answer: "I was very satisfied!" Somehow "very satisfied" probably falls very short of your true evaluation! Satisfaction is defined as "adequate or sufficient"--not exactly the gauge of customer devotion!
I am a huge believer in a 4 point scale. (No science just opinion.)
My experiences tell me that respondents are rarely ambivalent about a choice. If you force them to come down on one side of the fence or the other, they will think the extra second and give you their honest impressions.
I agree with Matt. 4 point scale (or a greater but still an even number) will force the user take a more precise position.
Answer This Question