A few weeks ago, my colleague John Jost of New York University circulated an email saying that he needed someone to be interviewed by a local television station. The task was to discuss and dispute claims made in two books about the futility of social psychology as a field of study. John couldn't do the interview himself because of other commitments, and I didn't volunteer because, well, I hadn't read the books. One of these books was "The rise and fall of social psychology: The use and misuse of the experimental method," by Augustine Brannigan (de Gruyter, 2004). I was intrigued by social psychology sounding like the Roman Empire or the Third Reich; so I bought the book.
Dr. Brannigan is a sociologist and he pulls no punches. He targets some of the most famous experiments conducted during the so-called "golden age" of social psychology. Many of his complaints are valid. Sherif's study on the formation of social norms was highly artificial; Milgram did not have a theory to explain obedient behavior until after he completed his studies; Zimbardo's prison study was not a true experiment at all.
Brannigan writes well and he has studied the original sources carefully. Yet, he is a zealot. The force of his conclusion is out of proportion to the strength of the evidence. His conclusion is that experimental social psychology has failed and must be abandoned [Note that this demand sets up any experimental social psychologist, who wants to defend the field, to be accused of bias. The defending social psychologist obviously does not want to be unemployed.]
The evidence is a critical review of certain "classic" studies. The criticism takes several forms: Lack of internal validity, lack of external validity, lack of theory, intrusion of moral judgments into scientific ones. Here's the rub. Brannigan's own argument is an evidence-based conclusion. If the evidence is bad, the conclusion is invalid. Brannigan does not consider the possibility that the studies he chooses to criticize may be a biased sample. His tactic is to imply that if the classic studies are bad, the rest has to be even worse. This conclusion may not be true. Classic studies are remembered because they were the original and most vivid demonstrations of a phenomenon (Brannigan agrees in most cases). Classic studies stimulate further research, which is often superior by methodological standards but less well remembered because it is considered ‘mop-up' work. Milgram's and Zimbardo's studies are exceptions because they are not being replicated. But then, the call to dissolve the discipline is moot if the discipline is not doing this sort of work any more anyway.
Brannigan protests too much. He notes correctly that social psychology has a built-in moral context. As ordinary people view their own behavior and the behavior of others primarily in moral (even moralistic) terms, a science studying this behavior must confront this context. By suggesting that we should retire to the disciplines of sociology and biology, Brannigan seems to say that moral questions are irrelevant on their turfs. Come again? Sociology [in particular] is disinterested in the moral implications of its subject? Refuting this claim by re-reading Plato or Marx would be too easy. Let's read Comte, Durkheim, and Weber again, shall we? The moral-contamination argument would require that we disband all social sciences that have anything interesting to say. What would be left is the kind of bean counting done in, say, certain demographical studies.
Another one of Brannigan's arguments is that social psychology is boring, trite, and redundant because all it does is replicate the insights of common-sense psychology. If folk psychology has already mapped out an understanding of self and social behavior, why do experiments?
The idea that folk psychology contains all we need to know about ourselves is seductive. The argument is twofold: (1) People do pretty well, by and large, navigating through a complex social world, thank you very much. (2) When experimental results are revealed, we often feel that the findings are obvious. Now, this is not enough to push experimental social psychology aside. In response to (1), we can note that people also often stumble, fail, and mess up (e.g., by trusting Bernie Madoff, or by having their racial attitudes polarized after hearing about the Gates/Crowley incident). In response to (2), we can note that the post-hoc sense of "this result is obvious" does not validate anything. The question is whether folk psychology can predict experimental results. Often it cannot, which brings me to the self-affirmation section of this post.
Folk psychology and some of its off-shoots (e.g., the so-called "self-esteem movement") assume that you can psyc yourself out of a funk and into a blessed state of high self-esteem by telling yourself good things about yourself. Whether this is auto-suggestion, self-hypnosis, self-persuasion, or simply perceptual priming is not the point here. The question is, does it work? According to Brannigan, there is no need to do a study because we already know that it works, thank you very much indeed!
But how do we know that we (meaning "most of us" presumably) already know? And how do we know that we know that we know? At this point, you need to imagine the sound of gnashing teeth (Brannigan's, that is). If you don't want an infinitely regressing argument, you must do empirical research.
In the latest issue of Psychological Science, Joanne Wood and her collaborators found that folk psychology only reaches a tentative consensus on the utility of positive self-statements. On a scale from 1 (strongly disagree) to 8 (strongly agree), the mean was 5.36 and the standard deviation was 1.68. This result means that 3 out of 10 participants did NOT believe positive self-statements are helpful. Wood's second preliminary finding was that a participant's own level of self-esteem predicted the strength of the belief in the efficacy of positive self-statements. This should give you pause.
The finding that self-esteem and belief in self-affirmation are correlated suggests two possibilities. One possibility is that self-affirmation works so fantastically well that people who use it end up with the highest self-esteem. The other possibility is that people who already have high self-esteem are more likely to use self-affirmation or that it is more likely to work for them than for people who start out with low self-esteem. Notice that folk psychology does not offer any answers here. You must do the experiment.
Wood and her collaborators did. They first measured participants' pre-study level of self-esteem. Then, they instructed them to say the phrase "I am a lovable person" 16 times over a 4-minute period. Then they measured participants' mood and momentary (state-like) level of self-esteem. Participants in the control condition did not self-affirm.
An armchair scholar who thinks he has a grip on folk psychology might predict that in the self-affirmation condition everyone's mood would improve and self-esteem would rise. A slightly more sophisticated armchair theoretician might predict that the positive effect of self-affirmation would be stronger for participants with low self-esteem than for participants with high self-esteem. The reason for this differential impact of self-affirmation could be a so-called "regression effect." When scores are low to begin with (low self-esteem), they have more room to grow.
Both these folksy ideas turned out to be wrong. Wood and her team found that participants who already had high self-esteem benefitted from self-affirmation, whereas participants who had low self-esteem were hurt by it. Self-affirmation, in other words, had a polarizing effect. Importantly, Wood and her team were able to argue that their results made good scientific sense. The point is, for experimental results to be credible, it is more important that they cohere with other theory and research than that they confirm common sense. In brief, Wood suggested that people with low self-esteem are harmed by self-affirmation because they just don't believe themselves to be lovable persons. Open declarative statements often trigger automatic counter-arguing. Someone who is overly self-critical and who says to herself "I am a lovable person," might spontaneously sneer at her clumsy attempt at self-indoctrination. As a result, self-esteem goes down even further. Instead of pulling the person's self-concept up, the positive self-statement now highlights the difference between the actual and the ideal self.
