We ran across an issue using a 59.94i sequence in Premier and had one of our editors run some tests. Here are his results:
If setting up a 29.97i sequence in premiere using 29.97i and 59.94i footage, premiere correctly treats each frame rate and there is no difference, actual or perceived. All fields are preserved.
If setting up a 59.94i sequence in premiere using 29.97i and 59.94i footage, things change. Premiere correctly handles the duration of the footage, but doesn't correctly handle the field order. The 29.97i footage is stripped of its fields making it look progressive on the broadcast monitor. The 59.94i footage maintains the fields (if viewed at 100% on the PGM monitor), but on the broadcast monitor it looks exactly like the 29.97i footage (it looks progressive). There seems to be no benefit from using a 59.94i sequence unless of course you are working with progressive footage and you want to maintain the actual progressive integrity beyond the perceived "feel." Does that make sense?
Also, please note that during these tests, the HARRIS VIDEOTEK RCU 1000 didn't "glitch" or switch monitoring modes (say from 29.97 to 59.94 or from interlaced to psf). It was locked onto 1080i/59.94 the whole time, also using internal reference.