Re: different scales(Likert and rank scales)



churmakov@xxxxxxxxxxx wrote:
Hi everyone! I have got a question that probably was answered already
somewhere. However, I could not find a simple answer. I have two set of

data, where the same products were rated by shop owners and customers.
All I want to do is to compare them... The problem is that Customers
ranked products i.e. they put them in the order of importance for them
( bread -10, milk - 9 , ... salt- 1). Shop owners answers are based on
5-point Likert scale... So is there any standard procedure to convert
them into one scale? Thank you very much for you time

No, there is no standard procedure to convert rankings and Likert-scale
ratings to a common scale. About the best you can do is correlate the
two. (Use Pearson or Kendall; Spearman is contraindicated by the small
number of categories, and the resulting large number of ties, in the
Likert data.)

.



Relevant Pages

  • different scales: compare?
    ... The problem is that Customers ... ranked products i.e. they put them in the order of importance for them ... 5-point Likert scale... ... So is there any standard procedure to convert ...
    (sci.stat.consult)
  • different scales(Likert and rank scales)
    ... The problem is that Customers ... ranked products i.e. they put them in the order of importance for them ... 5-point Likert scale... ... So is there any standard procedure to convert ...
    (sci.stat.edu)
  • Re: four-point Likert scale items
    ... ordered values as the measure of the construct with a Likert SCALE. ... consideration to quality of measurement at the design stage of a study. ... learned their statistics in psychology classes, ...
    (sci.stat.consult)
  • Re: Likert Scale Analysis
    ... additive scale. ... Jamieson uses "Likert scale" to refer to the ... "What approach should one take when specialist texts ... advantage of a parametric test may be negated if one or more of its ...
    (sci.stat.edu)
  • Re: four-point Likert scale items
    ... this article seems to confuse a single item with a set of ordered values as the measure of the construct with a Likert SCALE. ... So, the fact that the responses are at one end of the _response_ scale for one group, and at the other end for another group does not mean that the measurement is poor. ... learned their statistics in psychology classes, ...
    (sci.stat.consult)