Isolation transformers aren't usually very good, anyway (aside from galvanic -- shock -- protection).
For example, you can't use an isolation transformer (on the EUT or scope) to measure a high side switching circuit: the common mode signal (the switching itself) forces switching noise throughout all the cables in the system, ruining your measurement.
The switching current flows from the EUT, through its supply cable, across the isolation transformer's winding capacitance, up the scope's cable, and along the probe cable and ground clip. The entire loop has a complex reactive response, with a main LC component due to the loop inductance resonating with the isolation capacitance. The voltage drop produced along cables might not be a big deal, but the voltage drop across even a very small length of uncoupled ground, at the probe, is enough to ruin your measurement.
The magnitude of current flow (approximately Vsw(pp) / sqrt(L_loop / C_iso)) may also cause problems for the switch itself.
This is mainly a high frequency problem, but it still occurs at low frequencies, if the measurement is sensitive enough. Low capacitance isolation transformers are required for medical applications, to minimize ground leakage current to the patent (and to isolate other systems from possible defib transients!).
Tim