I based it on base model, and selected the same CPUID.
In other words, you can comparing CPUs not systems. That is not a system to system (apples to apples ) comparison.
the 3500 throws out features (besides the CPU) that the 5500 has. The Mac Pros don't.
As per the enclosure, for some, this could be an issue (3x 3.5" HDD bays vs. 4 in the MP). But it also includes an eSATA port, which the MP doesn't, so that sort of evens out IMO, as it can easily handle a backup drive/PM enclosure with multiple disks = not possible in the MP without consuming a PCIe slot.
If downsizing the cases didn't make a difference then why didn't Dell use the 5500 case with the different CPUID ? Just a tiny bit of common sense illustrates that this isn't "apples-to-apples" comparisons. That is standard backdoor for inserting the "Apple should sells mini or mid towers system that compete in the iMac price zone" arguments. Otherwise you keep that systems close to the same space.
CPU by themselves don't determine markets.