I would think that building the index on compressed data would take a bit longer due to the compress/decompress (higher cpu utilization).
The queries against them should take a comparable amount of time as against the data as other queries, allowing from slightly more time for decompression.
That is my theory based on some research. Definately not a definitive answer.
Keep in mind that data compression is not a performance enhancing option. Rather it is a resource conservation option. You need to perform some testing and compile a list of pros and cons to compare.