Himani Mishra I'd say it depends on what you are trying to do/find out with the results of your study. If, for example, one goal of your study is to compare the results to findings of previous studies, then it would perhaps be best not to change the scale for better comparability. If, on the other hand, the study purpose is mostly about examining relationships between different variables in your specific population, then it may not matter.

Himani Mishra

Yes, you may alter the scale points. It is often advised to maintain related scales, such as Likert-types, on the same point range. Yes, you may switch the scale points from one question to the next in the questionnaire.

The scale was strictly evaluated within a specific environment. This also applies to the sort of Likert scale you employ. Nobody will hold it against you if you convert a four or six-point scale to a five-point one.

One simple method for converting scale points is to multiply each scale point by two. The resulting new 10-point scale will only feature the five (even-numbered) scale points.

Keep it as is; it's fine if you use alternative point scales. It is determined by the type and characteristics of each inquiry. Consider it from the respondents' point of view. If the Likert scale is inconsistent, there should be a justification for this in your research objective.

Dear Himani,

I suggest that it would be possible for you to change the number of points on the Likert responses on the questionnaire with several considerations. The first point is that you would not be able to make a direct comparison of your results with studies that used the original questionnaire without completing an equating study as part of your research. The second point is that you would need to re-validate the questionnaire with the new response scales and recompute the reliability. That is a recommendation that I would make regardless of whether or not your changed the response scales as the validity of any tool depends on the population for which it was used. The final point is that I suggest that if you inform the originator of the questionnaire what you are doing as a courtesy.

Good luck,

J. McLean

Hi Himani Mishra

You can use different scale for different groups of population. If you want to compare findings of different population group, it would be better to use same likert scale for all population group. Another thing, you might consider reclassifying likert scale after collecting data. For example, you have likert scale of 1-5 for one population group and 1-7 for another population group. After collecting the the you can normalize the values on a scale of 0-1 and do the comparision. I hope this helps.

I usually go for seven points scale. Reading the inputs from the learned scholars, I had to thing over my methodology. Mr. Bhuiya has an important point to share when he says that the data can be standardized once data has been collected on different scale. Another point to share in response to what had been said by Respected McLean that the results are not comparable with the other studies. I would say that all the models are developed through standardized data so models are easily comparable without any problem regarding the number of points on scale.

Hi Himani,

My understanding is that you want to use scales from different sources in the literature which, unfortunately, are not in the same Likert range, but not to compare to some other past studies, but to answer your research question.

In this situation, you can adapt all scales to a 1-7 Likert range, as some people already answer at some point in this post.

In any case, you will have to perform reliability and validity tests once original validation does not mean you did not introduce some errors in the phrases due to translation or any jargon used in some specific research field.

Hope this helps.

Regards.

Why not use a 11-point scale (0 to 10, with 5 being the midpoint) with anchors? Most respondents can relate to 0 to 10 ratings. It was also uncomplicated to normalise these ratings for Multivariate Analysis.

cf attached paper