At ranges of a few hundred meters, will there be an appreciable difference in performance (ie range, bit rate) between 900 MHz and 433 MHz for digital applications in an urban environment?
Forgot to mention, I've read that some frequency ranges are more crowded than others, is that relevant here?
Also I live in the US.
433MHz will tend to be a bit better IMHO. It's the ability of the two frequencies to sneak between all the obstacles in our modern environment that is probably about equal but, at approximately half the frequency 433MHz will have more penetration due to the free-space link-loss formula. Link loss (attenuation in dB between transmitting antenna and receiving antenna) is: -
32.45dB + 20log\$_{10}\$(MHz) + 20log\$_{10}\$(kilometres)
At 500m, the link loss for 915MHz will be 32.45dB + 59.2dB + 0 = 91.7dB
At 500m, the link loss for 434MHz will be 32.45dB + 52.7dB + 0 = 85.2dB
Not much in it but 5.5 dB is better than a kick in the crotch!
In the US, the power limit for FHSS transmissions is 1W (30dBm) Non FHSS narrow band (3dB BW < 500KHz) is limited to -1dBm, wide band (3dB bandwidth >500KHz) is limited to 8dBm PSD, for 433MHz you can transmit at ~-5dBm, depends on your duty cycle. Also, the 900MHz band offers far more bandwidth, from 902 to 928MHz than the 433 band.
Another thing to consider, it is generally easier to find equipment for the 900MHz band, ie modems, antennae and various amplifiers than for the 433MHz band.
The 400MHz band has much lower power limits in most countries than the ISM bands at 900 MHz/2.4GHz/5GHz. So right off the bat 900MHz will be better.
If transmission powers are equal then it depends entirely on your structure. Are you propagating through outside walls to the inside? Between floors? Only on one floor?
Outside walls are usually concrete with steel re-enforcement. The steel rebar really attenuates RF signals and this makes propagation poor. The same can be said for propagating between floors. However on the same floor many office spaces have metal grids holding a false ceiling. This structure can be used as a successful waveguide and can help carry the signal around the building and through dry wall barriers. Polarization types can also make a big difference between poor propagation and good wave guide conduction. And in many general cases <800MHz the propagation isn't as good through these barriers.
Unless you are doing some sort of application specific install where you can optimize the antennas to the layout and structure of the building, all this info won't do you much good. If you're just looking for a generic solution that might work, then all you have to do is buy some cheap WIFI routers and do a site evaluation. Test them on 900MHz (new 802.11ah) or 2.4Ghz (802.11b and 802.11g) and even 5GHz (802.11a) in the areas you want to install.
EDIT:
I've done number of research projects in 100MHz/200MHz/400MHz/900MHz...5.8GHz for in building propagation. In general I've found that 900MHz tends to get better coverage than the lower frequencies (for the same power levels). I can't really share the details, but here's an example of the type of testing and simulation that many people have done to try to better understand in-building propagation. http://stakeholders.ofcom.org.uk/binaries/research/spectrum-research/in_building.pdf
In general the power limits on 400MHz isn't going to work well for you. We did several narrow band data comparisons and the performance at the legal power levels for all the ISM bands (and a few other bands like FM/VHF/etc.) is not very good for a variety of reasons. Go spread spectrum in one of the higher frequency ISM bands. But if you could tell us more about what you're trying to do it would help.
