4
$\begingroup$

Likert Scale has been widely in use for more than 80 years (it was published in 1932). However, the limitations of it are well-known:

  • The space between each choice is possibly not equidistant.
  • People keep endorsing the mid-point of the scale and avoid choosing the “extremes" options on the scale, even if an extreme choice would be the most accurate.
  • Attitudes of the population for one particular item in reality exist on a vast, multi-dimensional continuum. However, the Likert Scale is uni-dimensional and only gives 5-7 options of choice
  • Peoples’ answers will be influenced by previous questions

What is the latest development in Psychometrics? Is there any budding and promising replacement or advancement for Likert Scale?

$\endgroup$
  • $\begingroup$ The criticism you put forth here can be found in the Wikipedia article on the topic, however it is not obviously sourced over there (no inline citations). So [perhaps part of the question should be]: is this criticism actually based on empirical evidence? (Something that an answer here might choose to address as well.) $\endgroup$ – Fizz Nov 4 '17 at 6:35
  • $\begingroup$ And I'm not sure how your last bullet/point is relevant to the scale used, rather than to the overall structure [repetitiveness] of the questionnaire. $\endgroup$ – Fizz Nov 4 '17 at 6:41
  • $\begingroup$ It's not hard to agree least with the 1st and 3rd points you make; and others do as well ncbi.nlm.nih.gov/pmc/articles/PMC4833473 $\endgroup$ – Fizz Nov 4 '17 at 6:46
5
$\begingroup$

The visual analogue scale (VAS) has been used in some research. For a direct comparison in one study, see Vickers who compared VAS vs. 7-point Likert.

It's interesting that Vicker found a similar effect to that found in a highly cited paper by Dawes on 5/7-point Likert vs 10-point Likert, namely that more points on the scale result in a lower average score... despite the fact that the stuff measured in these two studies was completely different: muscle soreness and respectively customer satisfaction. Well, the probability of this coincidence is 50%, but I couldn't find any meta-analysis of this type of study.

One additional issue with Likert (and other ordinal) scales is that some researchers use the wrong statistics on it, like those assuming normal distribution, c.f. Bishop and Herron, who also discuss some finer points.

Refs (spelled out):

$\endgroup$
  • 1
    $\begingroup$ The papers you suggested in the comments are highly useful. $\endgroup$ – JetLag Nov 4 '17 at 7:39
  • $\begingroup$ I think that accept this response is negative, although the documents that quotes can help the user who asked. The 5-point scale is the most used for several reasons: it is from 5 response options when the variability works more powerfully, more response options introduce even more biases, it is obvious that many people are not used to handle a scale of 7 options. Nor is it commented on how psychometrics works to eliminate biases and questions raised in the development of the question. $\endgroup$ – hexadecimal Nov 4 '17 at 14:34

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.