It's a different kind of reliability. First of all, everything in the
Ancestral File is just a guide to help you determine the true facts.
Information submitted by individuals is no more accurate than the
submitter made it. That is why the submitter's name and address is
given: so you can contact that person and evaluate the data he has provided.
In the case of the LDS Medieval Study Group, you have data submitted by
scholars whose job is studying medieval records. That does not mean the
data is correct, only that it is their best estimate of the facts. I
suspect you can contact the Medieval Study Group just as you can any
other submitter and ask them about the reliability (ie, confidence) they
have in any particular piece of data. I suspect you will find that on
some of the data there is general consensus and other data there is a
difference of opinion amongst the scholars. It would be nice if they
could tack a number of 1 to 100 to each data point to indicate the
liklihood of its being correct.
The bottom line is that all data is fallible, and should be assumed to
be in error until you can convince yourself that it is correct
(sometimes you can do this simply based on the good reputation of the source).