Anecdotal information is problematic for several reasons, and I have highlighted some of these reasons in a speech I gave at the 2004 meeting of the Midwest Society for Pediatric Research.
Catherine Fiorello highlights the weakness of the anecdotal argument "I was spanked as a child and I turned out okay." Such an observation ignores the need for a comparison group. Such a group would help answer the question "What would I have been like if I hadn't been spanked as a child?" When large groups of both spanked and unspanked children are compared, the unspanked children tend to fare better. They are less aggressive and have fewer problems with depression.
This research is prospective, which is important. With retrospective data (how much did you spank your child in the last two years and how often did your child misbehave in the past two years) it would be impossible to disentangle causes and effects (did the child misbehave more because of being spanked or did the parent spank because the child misbehaved so often).
Common Myths of Children's Behavior. Fiorello CA. Skeptical Inquirer 25(3) (May/June 2001): 37-39,44.
Spanking by parents and subsequent antisocial behavior of children. Straus MA, Sugarman DB, Giles-Sims J. Arch Pediatr Adolesc Med. 1997 Aug;151(8):761-7. [Medline]
You can find an earlier version of this page on my original website.