Dear Dr. Albashayreh,

It is a controversial issue for using Likert scale with neutral point vs. not. If you selected a valid and reliable scale without the neutral option, then you have to use it as it is. However, you can pilot the scale and see possible comments/objections.

If you listed few items here, we may help better. It is possible prefer the number of steps based on the options and the stem of each item.

Good luck

When using a rating scale (e.g., Likert-type scale), two important choices, among others, are required: the number of response categories to be used, including the choice for an odd or even number of categories (with or without a midpoint), and the labeling of response categories (Weijters et al. 2010; Malhotra, 2007).

The decision to use or not to use midpoints in Likert-type scales is also culture specific. There have been a number studies that investigated the differences in response styles across cultures with reference to the midpoints (e.g., Lee et al., 2002; Chen et al., 1995; Harzing et al., 2012).

That is to say, you can change the response categories of adapted scale to fit to your research context with justification.

• Chen, C., Lee, S. & Stevenson, H.W. 1995. Response style and cross-cultural comparisons of rating scales among East Asian and North American students. Psychology Science 6: 170–175.
• Harzing, A. W., Köster, K. & Zhao, S. 2012. Response style differences in cross-national research. Management International Review 52(3): 341-363.
• Lee, J.W., Jones, P.S. Mineyama, Y. & Zhang,  X.E. 2002. Cultural differences in responses to a Likert Scale. Research in Nursing & Health 25(4): 295–296.
• Malhotra, N.K. 2007. Marketing research: An applied orientation. 5th Edition, Prentice-Hall of India, New Delhi.
• Weijters, B., Geuens, M. & Schillewaert, N. 2009. The proximity effect: The role of inter item distance on reverse-item bias. International Journal of Research in Marketing 26(1): 2−12.

Dear Dr. Albashayreh,

It is a controversial issue for using Likert scale with neutral point vs. not. If you selected a valid and reliable scale without the neutral option, then you have to use it as it is. However, you can pilot the scale and see possible comments/objections.

If you listed few items here, we may help better. It is possible prefer the number of steps based on the options and the stem of each item.

Good luck

this will remain a controversial issue ..  the investigator should decide about to include neutral response or not depends on the intended outcomes!

Hello there,

Intuitively, I would add 'neutral' answer category as why not giving a chance if someone is sitting 'on' the fence. Without this category we are forcing respondents to 'fall' to the right or to the left. On the other hand you are well secured with four categories as you may 'blame' references for that. 

Best wishes

Here is an article which addresses some of these issues, and which has a good literature review for your purposes: 

Armstrong, R. L. (1987). The midpoint on a five-point Likert-type scale. Perceptual and Motor Skills, 64(2), 359-362.

The authors found that there was little difference between 'neutral' and 'no answer' in the mind of respondents.

Hope this helps!

Dear Ala'a

with respect of the opinion and advises of the pioneer persons above my view is to use the scale as it built. Moreover, I advice to return to the author of the scale and get the justification of omitting the neutral answer, some authors made that in purpose.

I have attached a review article by two-well know experts in the field of item construction for surveys, which reviews the topics you mention.

One question I have is why you want to modify a scale that has worked well in the past. Unless the scale had problems with either reliability or validity, what would you gain?

Hello there,

I appreciate your contributions to this discussion. In addition to the review article suggested by Prof. David, I think the following articles are useful:

http://econtent.hogrefe.com/doi/pdf/10.1027/1614-2241.4.2.73

http://smr.sagepub.com/content/early/2013/12/02/0049124113509605.full.pdf+html

http://jpa.sagepub.com/content/32/7/663

In fact, the study is about role transition from education to practice among novice nurses. However, I'm exploring the possibility of adding a midpoint to one of the study tools because of the following reasons:

1. The tool will be used to measure comfort and confidence among novice nurses. I'm planning to use three tools that utilize a midpoint and I thought adding a "neutral" response to the fourth one would enhance the overall consistency of the questionnaire.

2. The target population is novice nurses who might be indifferent toward some tool items. I thought I should not force them to select the right "agree" or the left "disagree" responses.

3. I'm planning to go back to the respondents after applying an intervention and I definitely want them be satisfied with completing the questionnaire to enhance the response rate (before and after). Thus, I thought absence of a midpoint could discourage indifferent respondents. 

On the other hand, the tool has good psychometric properties and changing it might be risky because of the respondent's tendency to select the "neutral" response.

I think that keeping all of the scales on a consistent response format is a good enough justification for modifying the original 4-point scale. As far as can tell from the literature, the effect is not likely to large.

First of all, a measurement scale with good psychometric properties may not be changed unless situation warrants. Research context is important, therefore, previously validated measurement tools may require modification based on expert assessment or researchers' evaluation of the context. Having said that, if your research context requires so, as it seems from your statements, measurement tools may be modified, but a pilot test and validation is necessary. As David said, the effect is not likely to be large. For reference, you can look at the following reference:

• Leung, S. O. (2011). A comparison of psychometric properties and normality in 4-, 5-, 6-, and 11-point Likert scales. Journal of Social Service Research, 37(4), 412-421.

I concur with the responses provided here in regards to caution in altering a standardized tool.

As to the advantages of not  including a mid point, it helps to avoid the error of central tendency ( when participants  choose the "middle" choice as a default response pattern).  On the other hand, depending on the nature of the survey and the topic or variable it measures, a "neutral" option may be required to  reflect a valid point of view. As I tell my students: each choice has its pros and cons. The beauty of research is that you are in control and can make a choice that works best for your study, as long as you justify  it empirically.