The difference between 32 bit and 64 bit is that 64 bit generally runs faster, because you can handle more data at once. Also, it is the only way to make your computer recognize more than 3GB of RAM, unless you have an Intel motherboard that supports the /PAE switch. The reason 32 bit systems can't see the 4th gig of RAM is because the memory addresses that would normally point there are reserved for the BIOS. The /PAE switch, which is put in your boot.ini (only if running XP, I think), changes the way an Intel motherboard handles that last block of addresses. It redirects them to point at your RAM, enabling your machine to use that 4th gig.
The downside is that in 64 bit OSes, you lose the ability to run 16 bit programs, which is actually a huge problem. While most programs you run nowadays are 32 bit, almost all of them still use 16 bit installers. So if you buy a 32 bit program that says "64 bit compatible," it will run, but you might not actually be able to install it. In most cases, though the 64 bit OS will silently substitute a 32 bit installer in place of the 16 bit one, however, I have run into instances where it doesn't. (I was installing a $300 voice recognition package on a friend's 64 bit OS, and the silent 32 bit installer substitution failed, making it impossible to run that program.)
In theory, you can get around this problem by virtualizing a 32 bit OS and running the installer on that, but you'll need VMWare Tools in order to run it on the virtual machine, and set the install directory to the physical disk, not the virtual one.
I guess the short answer is that it's really up to you. It depends on whether or not you think the benefits are worth the risks. And before I forget, you should make sure your CPU supports a 64 bit instruction set. If it doesn't, you'll have to upgrade to one that does, else it won't work.