Will “the masses” always prefer a larger family tree to a better-supported family tree?

Genealogy and family history are not black-and-white sciences. We gather evidence and make a case for one version of the past or another, but the simple fact of the matter is you don’t know even recent facts like your own birthdate, and the evidence you have is not always be self-consistent or accurate. There are various models for how to evaluate the accuracy of genealogical research and/or its outcomes; for the purpose of this entry, however, we can simply observe that some research is more rigorous and some results better defended than others.

One method of evaluating the quality of genealogies is pretty well acknowledged not to work, and that is majority vote. The wisdom of the masses is a powerful thing… when the average members of the masses can be relied upon to be unbiased. But that is not the case in family history. Instead, there is a strong bias toward big trees.

This bias arises from several causes. There’s the desire to be connected to someone famous; the pleasing symmetry of a “‍full‍” tree; the jolt of happy brain chemicals upon finding a new person or missing link; the dissatisfaction of being left with an unanswered question, particularly one that relates in some way to one’s own identity; and so on. No matter the cause, there is an incentive to err on the side of a larger, fuller lineage than evidence can properly support.

And thus we see a proliferation of Big Bad Trees, by which I mean family histories containing more people than sources, more sources than rationales, and more leaps of optimism over fuzzy periods than walls of uncertainty documented with dozens of ambiguous or even contradictory records.

Is there any way to remove the Big Bad Tree incentive? Can it be reduced in novice researchers without quelling their enthusiasm? Anecdotally, some researchers are careful from their first day while others work for decades on ever bigger and badder trees; what is it that makes these two classes different?