Nathan A. Barclay wrote:

Straus and Mouradian's 1998 study divided mothers who spanked into three
categories depending on whether they "never," "sometimes," or "often"
spanked as a result of having "lost it." The outcomes for mothers who
sometimes spanked as a result of having "lost it" were significantly worse
than those for mothers who never spanked, and the outcomes for those who
often spanked as a result of having "lost it" were even worse still. But
the outcomes for mothers who never spanked as a result of having "lost it"
were very close to those for mothers who never spanked at all - slightly
worse, but either within the margin of error or too close to attach much
meaning in a study where self-selection bias is present. Note that that
study controlled for only one of several factors that I believe makes a
significant (if not huge) difference in how effective or dangerous spanking
is, yet it ended up with a group of spanking mothers with results extremely
close to the results of mothers who never spanked.

Also, it is important to note that Straus and Mouradian (1998) also
found that, among these mothers, the more non-cp used, the worse the
outcomes. In other words,
the non-cp methods were no better than spanking.

AF