Words from Diana Baumrind on corporal punishment...

A little background for you...

The World Wide Web is SO cool.

 I subscribe to a listserv called FAMLYSCI, which is sponsored by the National Council on Family Relations.  In a discussion on corporal punishment, Dr. Baumrind believed that her views had been misrepresented, and wrote to one of her colleagues about it. He posted her response to the list. I thought students in this course might be interested in seeing that researchers in developmental psychology are actual people (not just names in a textbook) and do feel very strongly about issues related to children. I asked Dr. Baumrind's permission to post her response for you to read, and she very kindly gave it. So, here it is.

(Note: CP is an abbreviation for corporal punishment.)

As she indicates below, Diana Baumrind asked me to pass her comments on to participants and onlookers re the recent corporal punishment debate.

Dear Bob:                                                               December 1, 1998

Thank you for sending me a portion of the important listserve debate on physical punishment and the role of values in science.  In the interest of truth and accuracy in reporting my views I would like you to share the following comments with the participants in the debate.

First let me say that I don't find it helpful to use such disjunctive categorizations as positivist or post-modern; extrinsic or intrinsic motivation; parent-centered or child-centered; pro-CP or anti-CP.  "Simple" disjunctions such as these are simplistic, and typically used pejoratively to put a "spin" on one's adversary's stance.  Thus the position its adherents would call "pro-choice" is called "pro-abortion" by those who call themselves "pro-life"; and the nuanced positions that either you (Bob) or I have on the subject of spanking are characterized as "pro-CP" and "pro-parent", because they are not "anti-CP" or "child-centered".

I would say instead that I am "family-centered".  In my view the true interests of parents and children are seldom antithetical-but in the few instances in which they are antithetical, parents are ethically obliged to put the interests of their dependent children before their own convenience. The dispute is about what constitutes the best interests of children and the family unit.

Deep value commitments should not be incompatible with the scientific method, which mandates objectivity, by which I mean truth and accuracy in reporting undeterred by personal prejudice.  When value commitments include (as Straus says his does) willingness to "ignore equivocal or inconsistent evidence" or to put a "spin" on one's representation of one's own findings then one's deep value commitments are indeed incompatible with objective science.  To quote Straus, when one "knows their theory is right" one "(up to a certain point) may ignore equivocal or inconsistent findings".  Why bother to collect data at all when one knows from the start one's theory is right?  Since data cannot speak for themselves, the public must trust that, in our role as scientists, we do not emulate political spin doctors by selectively reporting our findings or refusing to abandon pre-judgement when faced with "equivocal or inconsistent evidence".  Values may appropriately drive one's hypotheses and choice of problems, but empirical data must drive one's conclusions.

The "theory" Straus knows is right is that a parent under no circumstances should ever hit a child because even one instance increases the chances of harmful side effects.  But the risk analyses Straus conducts show no such thing and logically cannot, because they are probabilistic, not categorical.  Risk factor analyses apply to a statistical population, not to individuals. To support a categorical injunction against any use of disciplinary spanking Straus would have to show that the harmful outcomes he claims occur on average also occur in each case, or at least apply to each cell in the matrix.  The association between disciplinary spanking and each harmful child outcome could not be moderated by any such factors as age of child, cultural meaning to the child and parent of mild "CP", overall parenting style, or concomitant use of explanation and reason.  If associations were so moderated, Straus would have to qualify his conclusions by reference to the appropriate moderating factors.  Furthermore, the hypothesized moderating factors would have to be measured reliably and validly-or their possible effects (for example, of such covariates as parental warmth, reasoning, and nurturance) would remain unknown.  Since Straus's argument singles out CP, he would have to conduct similar analyses using alternative methods of punishment as independent variables to show that they have a less pernicious effect than "CP".  A proper test of Straus's blanket injunction theory requires pairwise contrasts on his 7-interval CP measure as well as tests for each of the theoretically meaningful cells in his matrix.  In view of the small effect size associated with his continuous measure of "CP" it is unlikely that Straus's blanket injunction against disciplinary use of mild spanking would be supported by such analyses.

Unlike smoking, spanking is a very weak and statistically unstable risk factor (in its positive association with the harmful outcomes Straus enumerates).  Also (unlike the parallel case with smoking) there is no consensus-scientific or otherwise-that "CP" is a generative cause of the negative child outcomes with which it is (rather weakly and inconsistently) associated.  Scientists know a great deal about the mediating processes by which smoking negatively affects the health of all smokers, and the effect size of smoking as a risk factor is huge by comparison with that of CP.  A generative theory indicting "CP" as uniquely harmful should explain why factors such as socialization context, developmental stage, meaning of "CP" to parent and child, and culture would not be expected to moderate the harmful effects of "CP" relative to its benefits, and why the predicted harmful effects would exceed those of all or most other forms of punishment.

Straus's deep value commitments do bias his judgement when they lead him to misrepresent or read the mind of his critics (e.g. Holloway, Bauman, Baumrind, Gunnoe, Larzerlere) as responding to a hidden agenda favorable to "CP," or in Larzelere's case to "Protestant fundamentalist" views, rather than to serious reservations about the analyses he uses and the logic by which he claims that these analyses support a blanket injunction against any use, by anyone, for any purpose, under any circumstance of mild disciplinary physical punishment.

I take Straus at his word when, contrary to Joan McCord who makes a theoretically coherent case against use of any punishment, he claims that it is only "CP" he opposes.  Straus takes the word of his survey respondents as to what they believe, but knows better than I what I really believe.  Straus characterized my position as "an ardent defender of CP" adding "She denies being a defender of CP".  That would make me a liar or unconscious of what I really believe.  Straus further misrepresents my position on "CP" when he states "In her 1996 article in Family Relations she repeatedly confounds no-CP with permissiveness" and "Most of the article is concerned with problems of permissiveness, not with problems of no-spanking".  In fact the confounding is entirely of Straus's creation since the focus of the Family Relations article neither claims to be, nor is, on either CP or permissiveness.  The focus of the article is on the broader issue of the role of the disciplinary encounter in child socialization, and in that context of punishment (including "CP") in the disciplinary encounter.  Since I report that almost all parents of preschoolers in our study did spank it follows logically as well as factually that (contrary to what Straus says I imply) "no-CP" is not synonymous with "permissive," and I do not suggest that it is.  As I report in that article, parenting styles cannot be differentiated by whether or not parents spank their preschoolers-Authoritative parents of preschoolers do spank more than Permissive parents and less than Authoritarian parents, but the contrasts are not statistically significant.  (Although I do not claim that "CP" is a necessary element in the Authoritative pattern we do not know whether the outcomes would be as successful-not only in producing compliance, but in generating agentic, prosocial behaviors, under the counterfactual condition that Authoritative parents used "no-CP").

For the record, I neither support nor reject the use of spanking as intrinsically immoral, but believe instead that aversive discipline and coercion should be used as little as needed and as effectively as possible, to assure a child's behavioral compliance with parents' legitimate and rationally defended directives. In my view it has not been established empirically that extrinsic reinforcers used appropriately undermine internal controls (see Cameron & Pierce; Nisan, 1992), but I prefer the relational perspective on compliance to the external control perspective.  Anyone familiar with me or my work knows that I do not value unquestioned obedience in a child or adult.

I do feel passionately about truth and accuracy in scientific reporting.  Biased reporting or "spin", or ignoring findings that contradict a theory one "knows is right", are practices not consistent with the scientific method.  Straus's polemical claims concerning the potential effects of physical punishment go well beyond his data, and in their categorical form cannot logically be supported by the particular analytic strategy he chooses to employ.  I would like to see data from the two national surveys Straus claims support his blanket anti-CP injunction reanalyzed by investigators less biased, or with contrary biases, using critical contrasts (e.g. between "No-CP" and "occasional CP"); theoretically composed cells in the matrix (e.g. differentiated by age, gender, ethnicity, frequency of use of "CP"); and parallel analyses with a variety of other disciplinary responses as independent variables.

I would also like to see a systematic study done of the adolescent and young adult outcomes following the "no-CP" laws in the Nordic countries.  Straus's prediction would be that youth of the generation who had experienced "no-CP" would be markedly less aggressive or depressed, and cognitively more advanced than the generation prior to the "no-CP" policy.  Although not generalizeable to societies such as ours that do not provide the same wonderful social support for parents as the Nordic countries do, evidence of a marked reduction in violence following the "no-CP" laws would go a long way to support Straus's position.  However, even if we grant that mild disciplinary spanking is risky behavior I suggest that, in the U.S., an effort at prevalence reduction (i.e. any use) by legal injunction would be less successful in dealing with its harmful consequences than educational strategies that focus on reducing the frequency of, or reliance upon, or inept use of "CP".

As you know, Bob, I have been funded to recode our longitudinal data to examine "CP" and other disciplinary strategies within a larger socialization context.  My tentative hypothesis is that appropriately used (especially within an authoritative context) disciplinary spanking is harmless relative to alternative forms of punishment.  Once our analyses are complete I will be better able to formulate an empirically based conclusion on what outcomes are associated with mild disciplinary spanking relative to other disciplinary tactics in our particular population.


Diana Baumrind

Click here to go on to the next topic (gender development) and for a few last words on parenting

Click here to send me email