Analyzing ordinal scale as interval-ratio

Hallo guys! I have a question: Have you even treated in your analysis an ordinal scale and specifically a Likert scale - Strongly Disagree, Disagree, Disagree a bit, Agree a bit, Agree, Strongly Agree (from 1-6) as an interval-ratio scale? if yes, how did you explain it in your Methodology Section.
And one more question: would it be a serious flaw to do the above in a paper? Thanking you in advance!

0 members like this

▶ Reply to This

Replies to This Discussion

Permalink Reply by Paul Donovan on November 9, 2009 at 1:54am

Γεια σου
Your Likert type scale includes colloquiallisms such as "agree a bit" which I haven't seen in the past. If I were to include it in the methodology section (which I think you should) I'd reference where I saw it used in the past. The Likert scales that are in common use are 5 point scales with "neither agree nor disagree" in the centre. In the case of this format, I doubt that you would have to refer to it in methodology other than to say that you were using a Likert scale. I hope this helps.
Paul

▶ Reply

Permalink Reply by Nikleia Eteokleous on November 11, 2009 at 7:23am

Γεια σου
Do you know any Greek? :) :) Thanks for your immediate response. But, I am still wondering if i can treat the Liker scale, which considers to be an ordinal scale, as interval-ration, that means conducting parametric statistics. Please let me know if you have any idea...thanks!:)

▶ Reply

Permalink Reply by Vincent A. DeBari on November 12, 2009 at 11:22am

Hi Paul.

I have a somewhat-related question:

Can total or overall scores from validated Likert-scale based survey instruments (eg, SF-36 or any of the various HRQL panels) be treated as interval data, In other words, if normally-distributed within the sample, can parametric statistical methods be applied to the total score? If so, might you have a reference for this?

Vince

▶ Reply

Permalink Reply by pavan kumar on November 9, 2009 at 2:28am

generally, liket scale will be like from 1-5, 1-7. most people prefer 5 point scale where in you have strongly disagree, disagree, nuetral, agree and strongly disagree. i have conducted few surveys using 5 point as well as 7 point scale.

you can do this on paper too...

▶ Reply

Permalink Reply by Nikleia Eteokleous on November 11, 2009 at 7:21am

Thanks for your reply! Just to clarify, you believe that i would be OK to go ahead and analyze it as an interval-ration scale conducting parametric statistics...Thanks :)