You might want to monitor the input level as well if you're using a cheap'n'cheerful generator. Some of the cheap ones have an output which varies quite a bit in amplitude as you adjust the frequency. Not a problem if you have half-decent lab equipment, of course.
As for errors, the cable resistance is unlikely to be a problem, but capacitance could be, depending upon the frequencies involved.
If this is a filter for audio frequencies where the value of C is relatively high then it's unlikely to be a problem.
If C is down in the picofarad range, however, such as would be found in some RF filters, then the lead and 'scope input capacitance could become significant.
An alternative to using a 'scope would be a VTVM or its modern transistor equivalent. Whatever you use as the measuring instrument, you need to make sure that the frequency response of the measuring equipment itself is satisfactory. This is another possible source of error. It's no good using a 10MHz bandwidth oscilloscope if you're trying to measure the response of a filter with a 3dB point set at 30MHz, for example.
P.S. Don't mean to be picky, and I know that pointing out spelling mistakes can be annoying, but it's "response" not "responce." No point losing marks unnecessarily, eh?
[This message has been edited by pauluk (edited 06-16-2004).]