Likert scale and PLS

This forum is closed, and read-only.
Locked
bommel
PLS Junior User
Posts: 2
Joined: Sun Dec 06, 2009 4:19 pm
Real name and title:

Likert scale and PLS

Post by bommel »

Hello,

Yesterday I used SmartPLS for the first time and it was really easy, but I still have one question / problem.

In my survey I used Likert scales and (I think) that such scales are not really metric. Other publications argue that Likert-scales are quasi-metric and as a consequence they can calculate the mean value or standard deviation (as you do it in SmartPLS in Bootstrapping-Total Effects for example).
My problem is, that it would be the best, to say in my thesis:

I use Likert scales and because of X who said that these scales are quasi-metric, I provide/use information (e.g. mean value) which is only reasonable for metric data.

But I don't know who explained this. As said before several publications write that Likert scales are quasi-metric but no one refers to another author or explains this.

Would be great if someone can help me!
Kind regards,
Dirk
User avatar
Diogenes
PLS Super-Expert
Posts: 899
Joined: Sat Oct 15, 2005 5:13 pm
Real name and title:
Location: São Paulo - BRAZIL
Contact:

Post by Diogenes »

Hi Dirk,

In fact the original idea was to use the summated rating scale to measure the LV, in this way the score will be almost metric.
Today everyone has used the items (ordinal) as if they were metric (as you said).

One article that I usually had quoted is STEVENS (1951), that was republished by WAINERMAN (1976).


“On the other hand, in favor of using ‘illegal’ statistics can invoke a kind of fundamental law: in many cases leads to fruitful results.”


STEVENS, S. S. (1951). Mathematics, Measurement and Psychophysics. In Handbook of Experimental Psychology, New York: Wiley, p.1-30.

WAINERMAN, C. H. (selección e introducción) (1976). Escalas de medición em ciências sociales. Buenos Aires: Ediciones Nueva Visión.

Best regards,

Bido
Locked