Which Likert scale should I prefer that of 1-4, 1-5 or 1-7?

You need to decide this based on the data you are looking for.

A scale of 1-4 does not allow the participant to give a neutral answer.  This may be desirable in some studies but not in others. 1-5 allows that neutral answer.  

A 1-7 scale gives TWO levels between "disagree strongly" and neutral (or whatever the values are labeled).  Do you need two?  Will the participants consistently be able to differentiate between 2 and 3?  Between 5 and 6?

I think that most often a scale of 1-5 works best, but as I said, it really depends on the details of your study and how meaningful the differences between scale values are.

Just as you said, there are advantages and disadvantages to each.  There's not a single "right" type of scale to use.  Just choose what seems to fit your study and be prepared to justify your choice as part of writing up your methods.  

I agree with both the responses. But in my personal experience, a scale of 1-5 has worked well. 

Following Q&A on RG was helpful for me to decide which scale to use for my research. Some of the answers have references to literature.

It might help you to decide what scale is best for you and find arguments to justify your choice.

Goor luck

I agree with all supporting 1-5, although there are believers that the neutral answer has to be counted as  negative. These are the supporters of 1-4. Lickert supported 1-5. Do we need to change it? 

I would recommend 1-7 with the end values only anchored (eg to strongly disagree and strongly agree). This is for two reasons. Firstly, people tend to answer slightly positively to please the question setter so there is no bias in including a middle value. Secondly, working memory modelling indicates that people can handle up to seven pieces of information at once so they should be able to discriminate between 7 choices. Only anchoring the end values also means you can treat the data as scale.

Thank you for the answers to my question. I'm overwhelmed with the uplifting feedback provided by you.

It seems that the most common is to use a scale of 5 with a neutral point and 2 positive choices and another 2 negative. But in line with the commentary of Peter, I have lately heard more tendency towards scale of 7. Questionnarie participants are my students and I believe that they will consistently be able to differentiate between 2 and 3 and between 5 and 6, as pointed Michael, but it forces me to not include many statements to evaluate. I'll keep thinking. Any suggestions are welcome.

Best regards!

I too agree that 4 point scale has obvious limitation. So, the obvious choice would be either a 5-point scale or a 7-point scale. In my own research projects, I have used both 5-point as well as 7-point scales in SEM models and both have produced decent results. So analysis wise i do not think it would make much of a difference and in literature we find evidence that supports use of both kinds of scales. 

Personally, i prefer 7-point scale as it surely gives more options to respondents for discriminating between 7 choices. 

Thank you for your comment, Anantjyoti. I agree with you. Finally, the scale used depends on how long the questionnaire.

Thank you for your comment, Harshvardhan. Now I have a better idea.

There is actually research on this topic, and I have attached a review article by two of the leading experts in the field that addresses this issue and several others related to item construction.

Dear David,

I note your comments.

Thank you so much for the paper you attached: I will read with great interest.

Best regards.

I would suggest you pre-test your questions using cognitive interviewing.  Also, don't be afraid to stray from the experts if your subjects need 3pt scale (elderly with dementia or very young) or use emoticons.  your challenge is help respondents not pay homage to past research.  The latter is good starting point.

Most studies concerning the use of either 5 or 7-item scales do not necessarily confer any absolute recommendation favoring one approach over the other. However, a study showed a slight support to use 7-point scale among respondents with more cognitive ability like student respondents, and use 5-point scale when respondents are general public (look at the reference 1 below). However, the authors seemed somewhat indifferent in using 5-point or 7-point scale at the end of their article. Another study conducted in European perspective utterly recommended 5-point scale to provide better quality of data compared to 7 or 11-point scale (look at the reference 2 below). Interestingly, the study falls in line with recommendations as outlined in reference 1, that 5-point scales be used when respondents are general public.

Hope they help you in deciding on this issue.

Overall I would agree with Dr. Rahman. I would also avoid even scales (1-4) if neutral response is important to the study. 1-7 spreads sentiment measurement a bit more accurately whereas 1-5 tends to group opinions.  It all depends on the diversity of response aggregates that will better address your research question or hypothesis.

Thank you very much for the last answers from Maria. Khandoker, Doug and Rimple

I'm sure they also help me a lot.

Hello Montserrat,

I would say the more points generally the better (I was often using 5 or 7 - it allows you to offer neutral answer). 

It gives people more flexibility in terms of their responses

Best wishes 

1 to 5 with  1and 2 being low, three being average/neutral and 4 and 5 being high.  Most respondents  can make this selection with ease.  Then ask for brief explanation of those who provide a 1 or 2 rating. 

Ok, Eugene, thank you!

Hi Witold.

I think that my final decision will be between scale of 5 or 7. When reading the documents that Khandoker attached kindly me, I will make the right decisión.

Thank you so much!

One issue that no one has discussed is how familiar your respondents are with this kind of question format. Based on another question that you have posted on RG, is it safe to assume that you will be working with students in higher education?

If so, I would recommend using 7-point scales. The main reason would be that they can yield more variance, and the only downside that I can think of is that they might be more difficult to answer for some types of respondents -- which would not apply here.

it seems,this based on the ability of the respondents and depends on the details of your study.I would recommend using 7-point scales.

Dear David and Hamid.

Thank you for your comments. Certainly they`ll help me in the final decision.

Indeed, my students are Higher Education and can distinguish scales 1-7 but my question was whether improved regarding scales 1-7 is really important or can be considered negligible.

Thank you!

Hello Montserrat, 

I see little if any issue when choosing 7 over 5 Likert type of answer categories. I vote for 7, but decision is all yours.

Best wishes

I like 1 to 5 because one can divide it more easily into three groups agree, disagree and neutral.   Going to 7 is like adding adding more choices in a supermarket.  It require a finer decision which may not have much psychological reliability  .  

I vote for 7-point scales. It better than 5 for separating in answers, and more accurate.Be sure the decision is yours.

Thank you very much, Witold, Eugene and Hamid for your answers.

I think the decisión will be taken when I pre-test my questions.

Regards!

Advantages:

Disadvantages:

Advantages:

Disadvantages:

Likert Scale 1-5 generally inherit the characteristics of Likert Scale 1-7 as mentioned above but Likert Scale 1-5's advantages & disadvantages are in between Likert Scale 1-4 & Likert Scale 1-7.

when you have homogeneous sample (like the case of students) having similar characteristics (e.g., educational level, age etc), its preferable to use 7 Likert scale to offer more variance.

Dear Colleagues.

First, what answer so interesting, clear and complete! Thank you for it, Han Ping Fung.

Second, thanks for warning me about the use of the 5 Likert scale in my case.

Regards!

I have used the five scale for decades with great results.  Much easier to develop a believable management analysis  with it. 

Thank you very much for your answers, Chalamala Srinivas and Eugene H. Fram. 

Regards!

1-4 scale does not gives you a neutral response so I think you should go with 1-5 scale as well as 1-7 scale.

Thank you Mohd Sadiq.

Sandip, selection of scale points depend upon your need. If you do not want to have a neutral point, you can go for 4 points (SA, A, SD, D or vice versa). At one end it pushes respondents to select one option instead of selecting neutral, on other it may be creating bias as well.

As far as I have seen, satisfaction is mostly measured with a neutral point such as SA, A, neither agree nor disagree, D, SD.

Decision is yours!

Thank you very much for your answer.

Kind regards!

Montse.

Selecting the scale depends on our choice to collect the data , for instance for me , I prefer to us five scale points A, SA,DA, SDA and N.

Hi,

Although there is a debate about what Likert scale to use in the questionnaires, I personally prefer 5 point Likert Scale comparing to 7 or 10 point Likert Scale.

According to Dawes (2008), with a Five - point scale, it is quite simple for the interviewer to read out the complete list of scale descriptors (‘1 equals strongly disagree, two equals disagree …’), and it is also quite simple to analyze the research data.

Reference: Dawes, J. (2008). Do data characteristics change according to the number of scale points used? An experiment using 5-point, 7-point and 10-point scales. International journal of market research, 50(1), 61-104.

Best

Thank you very much for your answers, Negussie and Abdullah.

Kind regards!

thank you

Does using a 4-point Likert scale weaken my questionnaire and, hence, my publication at the end?

I do not think that the use 4-point scale weakens the questionnaire rather, in my humble opinion, it strengthens it as it does make room for neutrality sometimes capitalise upon by some respondents to evade completing the questionnaire fully. Thanks and cheers.

Matthias Nkuda

Thank you @Matthias Nkuda!

Thank you very much, again, for the last answers (Haji Saediman, Lana Halaseh and Matthias Nkuda).

Best regards for everybody!

Montse.

While using a 4-point Likert scale is no longer the preferred approach, it certainly does not weaken your analysis to the point that it would prevent publication.

I have not seen or read a literature that disagreed with the used of 4 Likert point scale.

Thank you very much for wour answer, Kudirat Abiola and David L. Morgan.

Best regards!!

Using 4 point or 6 point likert scale won't weaken your research because it has less chance of getting a neutral response. For further clarification read - - Khan, polonsky, & Vincent (2016).

Thank you, Mohd Sadiq. Your answer is very important for me.

Best regards.

I think the scale that allows moderate or midpoint value is preferred, that is 5 or 7 points Likert scale. It also depends on what you want to achieve but I think the easy way out is 4 point in which the respondents are forced to make decisions.

Yes, Augustine Madu. It seems to be the most appropriate decision.

Thank you very much!

Best regards.

1-5

Thank you very much for your answer, Samir.

Best regards!

Montse.

When a 5 point Likert scale is used, there is the tendency for some respondents to select a 'neutral' or 'not sure' response to avoid making a real choice. However, with the 4 point scale, respondents are 'forced' to choose between favourable and unfavourable responses to each questionnaire item. This notwithstanding, to use a 5 or 4 point scale will always depend on what the reacher wants to achieve.

You are right, Nsoh.

Thank you very much!

Best regards.

Montse

Hi, a scale of 1-4 Likert is reported to distort outcomes. Also, it forces a choice when a respondent has no opinion and data on a Likert scale of 5 is more reliable than the data on the four scales. Further, respondents may not be reacting at all. In many cases it is better to know that they were neutral than not having them answer the question at all... On the other hand, 1-5 sometimes, is flawed, can't measure all attitudes to a problem or its findings may not be conclusive. Finally on 1-7 scale the answers of the participants may be influenced by prior questions... The decision is yours!

Thank you very much for your contribution, Nikos.

Obviously, the number of options in a question depends on the study you have in hand, what you want to get out of it ... But what is clear is that a scale that allows the respondent to show their indifference is better than causing them not to answer.

I have never favoured a 5-point scale. Why? I always believe that since Questionnaires employing such scale belong to the ordinal measurement level, how then will a neutral/undecided response be awarded 3points, greater than those who decided? Remember that in ordinal scale, numbers are assigned for identification and ranking. Hence, I always favour a 4-point scale, no room to 'sit on the fence'.

Oluwole's view corroborates my position on this matter. The essence of the questionnaire, in first place, was to get respondents' perceptions on the social phenomenon being investigated. Assuming all or majority of the respondents prefer neutral or undecided, what value will the instrument add to the research outcome if any. In line with Oluwole, the neutral or undecided option should attract a lesson value than diagree and strongly disagree. Research is about finding improved and better way to understand a given social reality or phenomenon. Thanks and cheers.

Very good reflection, Oluwole. Also thanks to you for your contribution, Matthias.

Thank you very much.

Regards.

While I understand Oluwole and Matthias's positions I suggest the positions are rooted in a bias towards Positivism. Maybe respondents prefer netral or undecided because the question is worded poorly? People choose neural or undecided when they don't have any other option and don't agree with the premis of the question. I agree that allowing a mid point/neutral response does not fit nicely with the underlying mathematical theory of an ordered response set that are equa distant. My suggestion. If you use a midpoint and the majority of respondents selected the mid-point then this social phenomenon should be investigated. I dont' think it is fair to say that 'no' social phenomenon exists. Another approach, use the mid-point to create an additional variable called based on the mid-point (i.e. Undecided). Can researcher argue that undecided not a real social phenomenon?

The basis for an ordinal scale is an "ordering" of the scores so that each one is greater than the next one. The mid-point of a scale is higher than the points below it and lower than those above it. Thus, it quite logically represents a point on a continuum of opinion, and what that means depends on how the question is worded.

I observe a number of typographical errors in my last submission. I wholly accept responsibility with apologies. Cheers.

Hello

1-7 is

Thank you very much, Doug, David, Matthias (don't worry, we all make mistakes) and Hamid.

Best regards.

1-7 is best.

Thank you, Selda.

Best regards.

Montse.

It is best to use a scale that covers many options (7)because that way the participant can select the option that best fits his or her situation. In addition, for test validation, scales of 3 and 4 do not work well and are problematic. Best wishes.

Yes, María Auxiliadora. Your answer has been of great help, although so have the answers that have opted for a smaller odd number of options or an even number of options to avoid intermediate positioning.

Thank you very much.

Best regards.

Montse.

7-point scales are a little better than 5-points—but not by much (Nunnally 1978). But, you should seek help from existing literature before forming your questionnaire.

If you are overwhelmed with low frequencies for some of the options, be practical. You can re-categorize the options depending on the closeness of the options (like what Jorge Ortiz Pinilla suggested here - https://www.researchgate.net/post/How_to_convert_or_deminish_a_7point_likert_scale_to_5point_likert_scale_in_SPSS).

Thank you, Mosahrop for your answer.

Best regards.

Montse.

I often use a 4-point Likert scale to get the specific response I need. In my opinion, neutrality does not justify something but then again it depends on your studies. For example, organizations should have discrimination policies; here is either you agree or you do not, being NEUTRAL won´t help me to further

the scale selection can be based on the types of respondent and also looking at the past and recent study. The advantages that we could do comparicon effectively.

Joy, thank you very much for your contribution. It has been a great help.