I think measuring wheels are much more susceptible to inaccuracy due to surface differences. The cal course I measured with my wheel was fairly rough asphalt. My wheel as plastic with a rubber strip around the circumference which is fairly sticky, but the wheel as a whole is pretty rigid. I don't think it handles the roughness of the asphalt well.
A few years ago I experimented with a couple wheels on a golf course. I measured out very short cal courses with a steel tape on two grass surfaces. One was regular looking grass, and the other was the grass on a fairway, which was very short and kind of spongy. The calibrations on each surface were very different, and they were also differences depending on how much I pushed down on the wheel as I rolled it. It did not inspire much confidence in measuring wheel measurements of cross country courses.
Also, many years ago I measured a track with a wheel. I first used my steel tape to measure from the 100-meter start line to the finish. It was spot on. Then I used that to calibrate the wheel. It gave measurements that were slightly different than 100 meters. I then measured the track following the line of the inside lane. Finally, I adjusted my measurement to account for the fact that tracks are designed to be 400 meters along a line 30cm (or 40cm if there is a curb) out from the inside line. The final measurement was just a few centimeters off. I actually think this is the best way to measure a track because it's easy to follow the inside line with a wheel. That would be very difficult on a bike. If the track is just part of a 5k then the riding error isn't going to matter much. But if the measurement is of just the track, the inevitable bike wobbling is going to make a big difference.