You know, that always bothered me. Because 5 isn't balanced evenly between 1 and 10. That's why I prefer to rate things on a scale from 0 to 10 (whether it be integers or rational numbers), b/c then 5 really is in the middle.
Benford's law might apply to the first digit of poll results; it would definitely apply if they reported total number of people estimated to be on each side, in populations of varying size. But if they are giving percentage results in close contests, the first digit would almost always be 4, and would never be 1 or 9.
But Benford's law doesn't apply to the last digit of numbers, which is what's at issue here.
My first instinct is always to pick 7, but I rejected that as obvious. For some reason, I didn't think of 5 -- 5 tends to turn up often in my life -- born in '55, both my names have five letters, etc. I like 4 because it's tidy, somehow -- I guess because of a square.
Considering that you didn't ask us to pick a random number - we could have picked our favorite number, for instance - there's no real reason for me to have attempted near-equal probabilities of picking each choice. Just habit. Experience says that following patterns is exploitable, and no reason to deviate from habit just because there's nothing to exploit. :)
Besides, I wouldn't want to offend any of the numbers by playing favorites...