Rationalists are far too quick to dismiss anecdotes as being "non-scientific." They are also too quick to accept all official-looking "data" that comes from scientists - when, in fact, anecdotes might sometimes be more trustworthy than scientific data.