I am curious what the modern accuracy for paleomagnetic azimuthal orientation measurements are. Let's say for a magnetite grain, is the orientation of the magnetic moment for a single crystal resolved down to the 1 degree? .01 degrees? or 0.00001 degree?
When core samples are taken in the field, their orientation must first be measured, generally with standard field tools such as a Brunton. These high quality, professional compasses are good to +/- 1 degree, which, additionally, are only as good as the stated declination for that day and location on Earth, which is probably around +/- 1 degrees. Add to that any errors in inclination caused by compaction and deformation and you have large uncertainties in predicted latitudes.
So, before the sample even makes it back to the paleomag lab, its orientation may already off by at least a few degrees, nevermind the inclination. This is why paleomagnetic data from a single sample is largely useless; many samples of the same age must be taken and a library of orientations built and compared to a global geomagnetic model. Even with that data, you may be looking at errors of up to 10 degrees.