I have a 6-point Likert scale for a psychology survey
consisting of the values 0, 1 , 2 ,3, 4, and 5
0 would be strongly disagree
between 2 and 3 would be neither agree nor disagree
and 5 would be strongly agree
Should I keep the middle option as "neither agree nor disagree"
or change it to "neutral"?
In other words, what is most accurate as the midpoint of a likert scale:
"neither agree nor disagree" OR "neutral"?
this is for social psychology attitudes towards social situations
Thank you
Hi Franklin,
1) from a statistical perspective there seems to be no strong reason to collapse two levels of a continuous scale, while keeping all others as they are. Why do you want to recode this? For a verbal presentation in front of an audience?
2) Your scale has no center/mid point, which can be desirable in some cases. Especially if participants tend to be undecided and are 'forced' to move to one side of the scale (left vs right side), instead of simply choosing the midpoint (50/50). In other words, if what participants see is:
strongly disagree < o o o o o o > strongly agree (mark the circle), then there simply is no 'neutral' option (is it would be the case for a 7 point scale).
3) I am not sure what Ashkan wants to say or where the analysis should go, but there are many ways to analyze your survey. A Likert Scale is most commonly used to treat a rating as a continuous measure. This allows some transformations of the data, for example calculating means and standard deviations (but there are other prerequisites for doing this, like normal distribution of the data etc). In this case, it simply makes no difference on whether you use 0 to 5 or 1 to 6. Neither is there a difference if you report the counts (not sums...) of each answer.
4) If you provide more details about what your goal is, you might get better answers :)
Best, René
Hi Franklin,
I agree with René with the lack of middle point but it might also be a problem since you want to add kind of a middle point. Since you're studying topics related to attitudes, neutral might imply a standing point but neither disagree nor agree sounds better. I also recommend the following readings.
- Krosnick, J. A., & Fabrigar, L. R. (1997). Designing rating scales for effective measurement in surveys. In L. Lyberg, P. P. Biemer, M. Collins, E. de Leeuw, C. S. Dippo, N. Schwarz, & D. Trewin (Eds.), Survey measurement and process quality. New York: Wiley.
- https://pprg.stanford.edu/wp-content/uploads/2010-Handbook-of-Survey-Research.pdf
Hope it helps.
Best,
İbrahim
René Schlegelmilch Thank you for your detailed and informative response, I appreciate your insights
Just so you know, I'm only in undergrad, so my background in statistics isn't strong
My professor decided to use the even likert scale, i'm just listening to him I guess
I've attached a screenshot of the likert scale options I've used
I'm following my prof's other examples
he put midpoint labels between the two middle values (in this case, since the values are 0,1,2,3,4, and 5, the two middle values are 2 and 3)
For example, as you see in this screenshot, I put "neither agree nor disagree" between the values of 2 and 3
But I want to be scientifically accurate.
Is it more scientifically accurate to put "neither agree nor disagree" or "neutral" between values 2 and 3?
Thank you,
Franklin
Hey there :)
I never use midpoint labels on a Likert Scale, just the endpoint labels, and participants never ask what the midpoints are for ... :)
So question back to you: Do you need this label? :) And again, if there is no midpoint, e.g. on a 6 point scale, then it is practically meaningless to give the "two points in the middle" the same label. This would be a 6-point version of a 5-point scale, which makes it even worse, because the most basic (methodological) assumption of a Likert scale is, that the 'intervals' between the labels are equally spaced (interval scale assumption). And collapsing two of the labels would deconstruct this assumption. So if you want an real option that says 'neutral' you might simply want to use a 5-point scale, for scientific accuracy. If you already have the data, then there simply is no 'neutral' option, because you have a 6-point scale.
Hope this helps.
Best wishes, René
I remember the main problem with such midpoint can be that some participants chose it when they don't want to answer the question/ or make decision. But it is questionable if is it right to code this "avoidance of answer" as a real "psychological midpoint". So, I'd prefer to use a scale without a midpoint.
Generally i prefer to use Likert scales with a mid point that is to say 3 5, 7 or 9 degree. In your case , i think, you have allready used a scale with 6 degree. Collapse the 2 central degrees will just lost informations.....so use your scale as it is with all your degree. It is acceptable to have no mid point. in this case it is thought that the mid point answer will be found more or less equally in the adjacent degree and some author prefer this way in order to minimize the central effect , the tendencu to use the central degree....
Francis
- Badges
- Science topic
- Similar topics
- Psychology