My opinion is that the value should largely be based on the traffic more than the population.
I own two city.com's that have populations around 35,000. The one domain receives MUCH more traffic. The domain that receives more traffic is a "free-standing" city v. being a suburb, the domain had formerly been used by the Chamber of Commerce, has a university of about 14,000 students and hosts training camps for several pro. sports teams. The other city.com is a suburb with next to no growth and very little traffic.
A lot of people will say base price on population. I also have a city.com of a very fast-growing town of 16,000. The domain gets 2-3 times more traffic than the city.com with a population of 35,000-40,000.
Also, don't assume a state.com is necessarily more valuable than a large city. Personally, I'd much rather have LasVegas.com or SanFrancisco.com than NorthDakota.com. I suspect the traffic is better and the leads more targeted than a state people don't travel to much.
Geo prices are all over the board-there's no formula that's going to make a lot of sense. Some of the sales of city.com's that I've seen reported are way above market and don't make sense for an investor.
With the three pure city.coms I've purchased this year, I've paid from 6 cents per person living in the city (population) up to 20 cents per person. The one I paid the most for was the best value and gives me a good return. On the other two, there's next to no return and won't be without some work/development.
That's my 2 cents.
Bill