4-point scale still valid?

[Roger Dingledine <arma@mit.edu>]

our list of options is:
Rank the following in terms of importance (low, moderately low,
moderately high, high):

Here are Karsten's original arguments for using a 4-point scale:

---

4. For preference questions, try to stick to a two or four level scale,
with no "indifferent" measure.  I strongly prefer a scale based on four
levels:  strong positive, mild positive, mild negative, strong negative
(usually given as agreement or preference).  The reason to exclude the
middle is to force the respondant to choose one way or the other.  More
levels don't give you higher accuracy (this comes from your sample size
-- it washes out), and dilute the statistical strength of the results.

---

I asked a political science friend her opinion on this, and she wrote:

---

1. I think it's perfectly fair, and more to the point, useful,
for a respondent to feel neutral about something. If they
truly don't give a shit, they should be able to say so. THis way
you won't be getting your weighting off. By forcing them to
choose one way or another, you are artificially inflating the
importance of that particular characteristic. Is the case white or
grey? I truly don't give a shit. It really doesn't matter. I'll
probably flip a coin. And that response will get bundled in
with the people who say that they prefer white if all other things
are equal, but not that they prioritize it. What's wrong with
apathy? You're not trying to motivate people; you're trying
to ascertain what they think about their computers.

2. Even if you accept his argument, it's not relevant to the scale.
He objects to having -2, -1, 0, 1, 2 because he doesn't like the
0 response. But you aren't asking that, you're asking 1 2 3 4,
all positive. You want to give people the opportunity to express
themselves, and that works better with an odd number. It gives them
a nice place to set the baseline to compare the important or
truly unimportant things to. I'd use a 5 or 7 point scale because
it's more sensitive. 2 is way way too small, and 4 is almost
too small.

---

Thoughts? Arguments? I'm beginning to think 4 is a bit small too.
--Roger