[Sigia-l] ballot usability redux

[Listera]

"Jared M. Spool" wrote:

> Actually, there have been. alphabetical listings and random order perform
> almost identically in most cases. David Fay and co. at Verizon Research
> have done a tremendous amount of work in Yellow Pages, finding that names
> listed at the top of the list have a significantly greater chance of being
> chosen than names at the bottom.
> 
> If the great state of California were really interested in making the list
> perform without any order bias, they'd make each ballot have it's own
> randomly unique order. (The beauty of 193 factorial is that it's a *really
> large number*.)

I don't think the great state of California has the wherewithal to produce
unique ballots by Oct. So it'll most likely be a choice between alphabetical
or random, as the news reports indicate.

The Yellow Book analogy is not complete. Voters presumably are *not* looking
to make arbitrary decisions based on position and convenience as you might
in a phone book (though this is California so you never know). They are
trying to spot a *specific* entry in an array of 193, most likely
distributed among multiple pages. Now, is it easier to find your *intended*
candidate among 192 other candidates when they are listed alphabetically or
randomly, over multiple pages?

Even though this is California, it's still an election. If we assume that
voters go into the booth having no idea who they'll vote for and make
completely arbitrary decisions when they first see the ballot (and therefore
we should randomize for that), then the question of whether elections mean
anything at all comes to mind.

> Maybe there should be a box on the ballot that says "I don't care. Just
> pick for me"

That may be unnecessary, with the California electorate.

Ziya
Nullius in Verba