I have a scale from 1-to-10 that currently reads like this:

10. Identical
9. Nearly
8. Extremely 
7. Very 
6. Quite 
5. Moderately 
4. Fairly 
3. Somewhat 
2. Slightly 
1. Barely 

I'm having problems with 3-4-5. The definitions of Somewhat, Fairly,
and Moderately are essentially identical. Although the numbers are
rather clear, I wonder if someone can help clarify the language and
suggest better words to use for this scale?

Thanks!

---

Request for Question Clarification by justaskscott-ga on 01 Jun 2005 14:16 PDT

What is being rated?  It sounds like the scale is being used to
compare one thing to another thing.  What are these things?

---

Clarification of Question by braitman-ga on 01 Jun 2005 15:00 PDT

I didn't think it important to mention what is being compared, since
I'm only trying to get terms of comparison along a scale. FYI, what is
being compared is music, songs similar to each other. But it could
just as well be automobiles or fish. Yes?

Thanks!

Hi Braitman,
Your question demonstrates why on questionaires such scales are
usually just given in numbers with a definition for the extreme
choices:  1 = no similarity,
10 = identical.   It also demonstrates why such scales seldom offer so
many choices.  Unless the subject provides some objective criteria,
different persons will not answer the same way when responding, thus
the results of the questionaires will reflect subjective personal
impressions, especially if a scale of words is provided, even if the
words are only intended as an explanation of the 1 to 10 scale.
As you point out, "somewhat", "fairly", and "moderately" all mean
about the same thing, but individuals, when referring to the table
will be influenced by their own perceptions of the words' meanings,
which will influence their response on the questionaire, just as my
own response would be influenced by my feeling that "extremely" is
closer to "identical" than "nearly" is.

I know my own feeling when filling in such a questionaire (hotel
service, for example) with more than five levels is to simplify it. 
After puzzling over the finer differentiation for the first couple of
items, I give up on using a couple of the choices, often on the best
classification, because with so many choices  nothing can be perfect 
- or in your list, "identical".
Unless you tell the participants that there are some identical pairs
in the excercise, they will probably be very reluctant to pick that
choice, assuming that there is possibly some subtle difference that
they did not recognize.

It could be, of course, that this is just what you want to test, the
differences in individuals' responses when comparing similar songs,
and comparing theirs with your own more objectively prepared table of
"correct" responses.
Here again, I think, providing the words would confuse rather than
improve the responses.

(Prior to the previous paragraph, I was assuming that you had no
preconception of "correct" responses and would just be looking at the
results from the participants' in relation to themselves.)

If that is not helpful in any way, sorry, it is only a comment and no
criticism of your project - which sounds intriguing.
Myoarin

I agree that there is a reason that GMAT tests, personality
questionnaires, etc. use 4-5 ratings groups, rather than 10.

I think the bast would be:
Very Dissimilar
Slightly Similar
Moderately Similar
Similar
Almost Identical

I also like when there are only 4 choices, so there is no
middle-of-the-road answer. This forces people to be left or right of
center on every decision, which makes them think about their answers
more. At least if does for me.

Cheers,
_Dave