Because they confuse Likert items with Likert scales, many authors look at a single Likert item
and conclude that the 1 2 3 4 and 5 options form an ordinal scale at best, and therefore data based
on these scales must be analyzed as though they are ordinal. I have two responses to that form of
“logic”.
When your read that MacArthur graduated first in the West Point class of 1903, that means he
was at the top of his class ahead of whoever was second, third, fourth, fifth, etc. What is it about
any Likert item 1 2 3 4 5 (much less an Likert scale) that can be expressed in ordinal numbers? Is
strongly agree fifth, ahead of agree at fourth, and neutral at third, disagree at 2nd, and strongly
disagree at 1st? This doesn’t make sense, even at the Likert item level, much less at the Likert scale
level.