# When is it okay not to use “middle” or “neutral” option for Likert scale?

I have couple of questions. Q1: How important is it to have a "Neutral" option in Likert scale. Does it depend on the queries asked? Or is it always preferable to have a "middle" option. The context of this question is that a questionnaire study was conducted in which passengers of a new bus system in the city were asked several questions regarding their "level of satisfaction" of various attributes. The actual survey had 5 options: (i) highly satisfied, (ii) adequately satisfied (iii) satisfied (iv) dissatisfied & (v) highly dissatisfied. It does seem that it is skewed more towards "satisfied" option; however having a "neutral" option would also imply that those users are "okay" with the system, thus making them move towards "satisfied group". So, Q2: In this scenario, should "adequately satisfied and "satisfied" be combined as one category since they essentially tell the same thing. OR should the "satisfied" option be treated as "neutral" since they're essentially saying "we're okay with it"?? Thank you in advance.

I have couple of questions. Q1: How important is it to have a "Neutral" option in Likert scale. Does it depend on the queries asked? Or is it always preferable to have a "middle" option.

If you do not have the "neutral" option, you force your responders to state their preference. The "neutral" option enables them not to have opinion. If people are forced to make a choice you end up with some answer, on another hand, if people feel that they don't have an opinion they may not answer your questions, or provide random, rubbish answers. There is lots of examples of surveys and questionnaires where the "neutral" option was, or was not included. It is up to you.

(...) however having a "neutral" option would also imply that those users are "okay" with the system, thus making them move towards "satisfied group".

The scale you quote is strange--are you sure that the options are correct? Nonetheless, the "neutral" option should be neutral. If it is not then you can get rubbish results. Creating asymmetric scales does not sound reasonable. I guess, many people would not read the options carefully and treat the middle option as "in the middle" rather then as "satisfied". By creating asymmetric scales you are playing with fire and you could come into issues with interpretability of your results.

So, Q2: In this scenario, should "adequately satisfied and "satisfied" be combined as one category since they essentially tell the same thing. OR should the "satisfied" option be treated as "neutral" since they're essentially saying "we're okay with it"??

How you aggregate and present your results is up to you and depends on your aims. Any way of aggregating may make sense in different scenarios. For example, you could as well aggregate everything other then "very dissatisfied" if you want to talk about people who are not "very dissatisfied" etc.

• Thank you Tim. It is evident to be careful in using asymmetric scales. Just to clarify, the questions were asked by the survey team "in person" from respondents rather than handing them the survey forms. A pilot study was conducted beforehand in which the middle option was "yeah, we're okay". I think since the native language is not English here that's why the middle option was essentially addressing "okay" rather than "no opinion". Your answer has highlighted some important aspects nevertheless. An extra care must be taken in written options even when asked in person. Thank you again – M COFF Aug 1 '17 at 1:49
• I think "neutral" and "no opinion" are different things. Neutral should indicate intermediate satisfaction. No opinion may indicate not taking buses, give a damn, did not understand, ... If you offer a neutral option and people use it in the sense of "no opinion", then you should not treat it as something in between "slightly yes" and "slightly no" (i.e., ordinal or even continuous). – hplieninger Aug 1 '17 at 12:00
• @hplieninger agree, this is what I meant. "No opinion" = "neither agree, nor disagree". This depends on wording of actual question and available answer options. – Tim Aug 1 '17 at 13:07
• Well, I'd rather say "No opinion" $\ne$ "neither agree, nor disagree". But I agree that it depends on context. – hplieninger Aug 1 '17 at 13:11

The issue of the middle (neutral) value in Likert scale has a very long history and people will argue for both sides as whether a middle value should be included or not. I can probably find an equal number of articles that support both sides. In the end the structure of your Likert Scale depends on how you can justify it based on the research context and that you can adequately defend. Personally, I see the middle value as being an easy (if not abused) option by respondents. Most of the Likert surveys that I undertake do not use a middle value. The analysis of the responses can infer whether there is a valid neutral view, within some degree of confidence.

More of an issue is the symmetry of your Likert scale. I agree with Tim's reply that the "satisfied" option should be treated as "neutral" since that is what you are inferring with your scale.