Size-Biased Sampling

A way of sampling where an item's chance of being picked grows with its size or its number of connections, so larger or more-connected items are over-represented. It is why a name occurrence pooled from every friend list looks more popular than a random person: well-connected people appear in many lists, so they get drawn more often. The same effect makes classes feel bigger when you ask students than when you ask the registrar, because more students sit in the large classes.