Abstract :
Educattion researches in surveying researches of qualifying type are generally exposed to
Nominal or Ordinal classified levels. In ordinal level in opinion polling papers, most
researchers deal with choices such as very good (very much), good (much), OK (to some
extent), poor (little), very poor (very little) and so on. Researchers are forced to analyze the
data in order to correspond them to quantity. For this action, the 5-level Likert spectrum
(American Sociologist) is used and correspond them to 1, 2, 3, 4, 5 with equal distances of
same value, so it used non-parameter tests because of non-parameter data. Since the
respondents’ view was considered as a kind of show-off, itʹs obvious that, first, that is a
difference between beliefs and views which lead to difficulty in analysis. Second, these is a
kind of ambiguity and overlapping in the vicinity of the border of choices that can lead to
error in deviation, variance, and test use. This problem also increases the analysis problem
and reduces the precision. In this paper, the purpose of writer is to have a fuzzy approach
to Likert spectrum and use the fuzzy model for research analysis.
