Why is providing a numerical rating (say, 1-10) better than a textual
or verbal rating alone? I can think of a couple of reasons:
- it's easy to interpret
- it's easy to compare different people's opinions
- it forces people to make a decision
- it can factor directly into a decision

But there are probably more, better reasons. Has there been any
research on this? Are there well-accepted reasons why numeric ratings
are good? Or bad, even?

---

Request for Question Clarification by czh-ga on 31 Aug 2005 15:43 PDT

Hello craignm-ga,

Choosing the best feedback mechanism or rating scale will depend on
what you're trying to measure and what you hope to learn from the
feedback. Numerical ratings will be appropriate for some situations
while verbal comments will provide much more useful information in
others. Please tell us more about the context of your question so that
we can find you the information you need. Thanks.

~ czh ~

First of all, let's look at the various kinds of numerical scales we have.

1) Ordinal scale: This has to do with ranking the extent to which a
certain attribute is present (such as a classroom rank for students,
or the order in which participants finished a race). So 1st and 2nd
might be separated by a teeny bit, but 2nd and 3rd by a huge amount.
Also, there can be no zero-eth rank.

2) Interval scale: Each number here represents an actual amount and
the difference between two consecutive numbers is fixed. A zero is
present in this scale, but it's not a "true" zero. For example, the
temperature scale or an intelligence scale. An IQ score of zero or a
temperature of zero degrees does not mean that intelligence and
temperature do not exist at all.

3) Ratio scales: Here, we're measuring the actual amount of something.
For instance, 4 litres of water means, there's 4 actual litres of
water, and 0 litres means there's no water at all. The zero has its
"true" meaning.

Now let's compare this to a non-numerical scale.

This would be the Nominal (name) Scale. It indicates no amounts and is
used for identification or classification purposes only. If two
entities have the same number associated with them, they belong to the
same category, and that is the only significance that they have. So,
male-female, black-blue, 1-2...all these are just names. Numbers mean
nothing...they're just names too, like the numbers on a football
jersey.

Well-accepted limitations of the Nominal (non-numerical scale):
(1) Only a frequency count is possible in terms of mathematical and
statistical analysis. And arbitrary conversion to a numerical system
is anyway needed for this (example, by "renaming" Males as 1 and
Females as 2 to analyse the results of a demographics survey).

(2) Much lesser sensitivity and precision than a numerical scale.

(3) Should not be used to measure behaviours/activities/phenomena that
can be quantified.

(4) Categories on a nominal scale MUST be mutually exclusive, that is,
either-or. Either you are male, or you are female. Either the colour
is brown, or it is black...shades in the middle are not possible on a
non-numerical scale.

These points are besides the ones you mentioned yourself. As you can
see, numerical scales provide you with a lot more options. Of course,
if you're just collecting demographic data, a nominal scale is
perfectly fine. But otherwise, you're better off going with a
numerical scale.

Is there any academic research that you can point to that discusses
the advantages or disadvantages of numerical scales?