ATTITUDES AND SOCIAL COGNITION
Establishing a Causal Chain: Why Experiments Are Often More Effective
Than Mediational Analyses in Examining Psychological Processes
Steven J. Spencer, Mark P. Zanna, and Geoffrey T. Fong
University of Waterloo
The authors propose that experiments that utilize mediational analyses as suggested by R. M. Baron and
D. A. Kenny (1986) are overused and sometimes improperly held up as necessary for a good social
psychological paper. The authors argue that when it is easy to manipulate and measure a proposed
psychological process that a series of experiments that demonstrates the proposed causal chain is
superior. They further argue that when it is easy to manipulate a proposed psychological process but
difficult to measure it that designs that examine underlying process by utilizing moderation can be
effective. It is only when measurement of a proposed psychological process is easy and manipulation of
it is difficult that designs that rely on mediational analyses should be preferred, and even in these
situations careful consideration should be given to the limiting factors of such designs.
As the field of social psychology has developed, researchers
have increasingly turned to the examination of psychological pro-
cesses. This development is perhaps inevitable as the field ma-
tures—fewer new phenomena are being discovered and increasing
attention is being paid to understanding known findings. We firmly
believe that understanding psychological processes is fundamental
to advancing the field. Nevertheless, over the years we have
become concerned with what we see as an overemphasis on one
particular way of examining psychological processes: regression
models based on a seminal paper by Baron and Kenny (1986). To
be sure such analyses have their proper place—and we plan to
describe it. However, it is our contention that this analysis strategy
is overused and has perhaps been elevated as the gold standard of
tests of psychological processes and may even be seen in some
quarters as the only legitimate way to examine them. The aim of
this essay is to point to broader ways of understanding psycholog-
ical processes in experimental contexts and to highlight the
strengths and weaknesses of various approaches.
The study of psychological processes has a long history in social
psychology. Most classic theories in social psychology theories
have compelling accounts of such processes. For example, cogni-
tive dissonance theory (Festinger, 1957) proposes that when peo-
ple have two thoughts that are psychologically inconsistent, they
experience an aversive arousal that motivates them to change one
of the cognitions. Clearly, the aversive arousal is the mediating
psychological process in this account. In those early days, social
psychologists typically tested their theories by demonstrating an
effect and then made plausible arguments about psychological
processes. Much of the time these arguments were not even tested.
For example, Festinger and Carlsmith (1959) demonstrated that
people paid $1 to lie about a boring task by saying it was fun came
to believe their lie, whereas those paid $20 to lie did not. They
argued that people in the $1 condition changed their attitude
because the inconsistency between what they believed about the
task (i.e., that it was boring) and what they said about the task (i.e.,
that it was fun) created an aversive state of arousal—cognitive
dissonance. They also argued that people were motivated to alle-
viate this aversive state, and because it was easier to change one’s
beliefs than to take back one’s actions, people resolved the incon-
sistency by changing their beliefs about the task (i.e., they thought
the task was fun).
As the field developed, it became more common for statistical
evidence to be offered for such hypothesized accounts of psycho-
logical processes. For example, research on the elaboration likeli-
hood model of persuasion (Petty & Cacioppo, 1986) began to
include measures of thought listing as an indicator of elaboration,
and research on group polarization began to examine the number
of arguments that were made in a discussion to examine this factor
as an indicator of process (Burnstein & Vinokur, 1977). These
authors demonstrated that a factor produced an effect and then they
measured the proposed psychological process and demonstrated
that the process occurred more often when the effect occurred and
that the measurement of the process correlated with the measure-
ment of the effect. For example, Burnstein and Vinokur (1977)
Steven J. Spencer, Mark P. Zanna, and Geoffrey T. Fong, Psychology
Department, University of Waterloo, Waterloo, Ontario, Canada.
The writing of this article was supported by grants from the Social
Sciences and Humanities Research Council of Canada and the Canadian
Institute of Health Research. We thank Joanne Wood and Judy Harack-
iewicz for their helpful comments that provided much needed clarification
of our arguments. In addition (without implying that they agree with our
analysis), we thank Reuben Baron, David Kenny, Lee Fabrigar, and four
anonymous reviewers for their generous comments on an earlier version of
Correspondence concerning this article should be addressed to Steven J.
Spencer, 200 University Avenue West, Psychology Department, University
Journal of Personality and Social Psychology, 2005, Vol. 89, No. 6, 845–851
Copyright 2005 by the American Psychological Association 0022-3514/05/$12.00 DOI: 10.1037/0022-35184.108.40.2065
signs have been overused and the Baron and Kenny (1986) type of
analysis often has been misapplied.
Finally, we believe that it is important to keep psychological
processes in their proper perspective. In our view, any healthy
science should have room for discovery and the development of
new theories as well as explanation and the refinement of existing
theories. Theoretical arguments about psychological processes by
their very nature tend to emphasize explanation and the refinement
of existing theories. In evaluating the theoretical contribution of
theories, let us rightly value the place of psychological processes
in social psychology, but let us also value the discovery of new
phenomena and the development of new theories.
In addition to their theoretical contribution, the practical contri-
bution of theoretical accounts of psychological processes should
also be considered. At their worst, accounts of psychological
processes (we will not even say theoretical accounts here) can
become an infinite regress of ever-finer intervening causes and the
worst sort of navel gazing. For example, some accounts of psy-
chological process propose mediators that are conceptually very
similar—perhaps indistinct—from either the independent variable
or dependent variable being examined. It may be easier to find
statistical evidence for such accounts, but their theoretical and
practical significance is suspect. At their best, however, theoretical
accounts of psychological process can provide important insights
that allow us to intervene to make the world a better place. It is
such psychological processes that should be valued, and a number
of different experimental approaches should be seen as valid ways
of testing such ideas. What we propose is that people consider a
broad range of possible designs when examining psychological
processes and select the design that is best suited to the particular
problem being studied.
Baron, R. M., & Kenny, D. A. (1986). The moderator–mediator variable
distinction in social psychological research: Conceptual, strategic, and
statistical considerations. Journal of Personality and Social Psychology,
Bollen, K. A., & Stine, R. (1990). Direct and indirect effects: Classical and
bootstrap estimates of variability. In C. C. Clogg (Ed.), Sociological
methodology (pp. 115–140). Oxford: Blackwell.
Burnstein, E., & Vinokur, A. (1977). Persuasive argumentation and social
comparison as determinants of attitude polarization. Journal of Experi-
mental Social Psychology, 13, 315–332.
Cohen, J., Cohen, P., West, S. G., & Aiken, L. S. (2003). Applied multiple
regression/correlation analysis for the behavioral sciences (3rd ed.).
Mahwah, NJ: Erlbaum.
Davies, P. G., Spencer, S. J., Quinn, D. M., & Gerhardstein, R. (2002). All
consuming images: How demeaning commercials that elicit stereotype
threat can restrain women academically and professionally. Personality
and Social Psychology Bulletin, 28, 1615–1628.
Davies, P. G., Spencer, S. J., & Steele, C. M. (2005). Clearing the air:
Identity safety moderates the effects of stereotype threat on women’s
leadership aspirations. Journal of Personality and Social Psychology,
Fein, S., & Spencer, S. J. (1997). Prejudice as self-image maintenance:
Affirming the self through negative evaluations of others. Journal of
Personality and Social Psychology, 73, 31–44.
Festinger, L. (1957). A theory of cognitive dissonance. Stanford, CA:
Stanford University Press.
Festinger, L., & Carlsmith, J. M. (1959). Cognitive consequences of forced
compliance. Journal of Abnormal and Social Psychology, 58, 203–210.
Harackiewicz, J. M., Abrahams, S., & Wageman, R. (1987). Performance
evaluation and intrinsic motivation: The effects of evaluative focus,
rewards, and achievement orientation. Journal of Personality and Social
Psychology, 53, 1015–1023.
Judd, C. M., & Kenny, D. A. (1981). Estimating the effects of social
interventions. Cambridge, England: Cambridge University Press.
Judd, C. M., Kenny, D. A., & McClelland, G. H. (2001). Estimating and
testing mediation and moderation in within-subject designs. Psycholog-
ical Methods, 6, 115–134.
Kenny, D. A., Kashy, D. A., & Bolger, N. (1998). Data analysis in social
psychology. In D. T. Gilbert, S. T. Fiske, & G. Lindzey (Eds.), The
handbook of social psychology. (Vol. 1, 4th ed., pp. 233–265). New
Kmenta, J. (1997). Elements of econometrics (2nd ed.). Ann Arbor, MI:
University of Michigan Press.
MacKinnon, D. P., Lockwood, C. M., Hoffman, J. M., West, S. G., &
Sheets, V. (2002). A comparison of methods to test mediation and other
intervening variable effects. Psychological Methods, 7, 83–104.
Muller, D., Judd, C. M., & Yzerbyt, V. Y. (2005). When mediation is
moderated and moderation is mediated. Journal of Personality and
Social Psychology, 89, 852–863.
Petty, R. E., & Cacioppo, J. T. (1986). Communication and persuasion:
Central and peripheral routes to persuasion. New York: Springer-
Quin ˜ones-Vidal, E., Lo ´pez-Garcı ´a, J. J., Pen ˜aranda-Ortega, M., & Tortosa-
Gil, F. (2004). The nature of social and personality psychology as
reflected in JPSP, 1965–2000. Journal of Personality and Social Psy-
chology, 86, 435–452.
Shrout, P. E., & Bolger, N. (2002). Mediation in experimental and non-
experimental studies: New procedures and recommendations. Psycho-
logical Methods, 7, 422–445.
Sigall, H., & Mills, J. (1998). Measures of independent variables and
mediators are useful in social psychology experiments: But are they
necessary? Personality and Social Psychology Review, 2, 218–226.
Son Hing, L. S., Li, W., & Zanna, M. P. (2002). Inducing hypocrisy to
reduce prejudicial responses among aversive racists. Journal of Exper-
imental Social Psychology, 38, 71–78.
Storms, M. D., & Nisbett, R. E. (1970). Insomnia and the attribution
process. Journal of Personality and Social Psychology, 16, 319–328.
Wegener, D. T., & Fabrigar, L. R. (2000). Analysis and design for non-
experimental data: Addressing causal and noncausal hypothesis. In H. T.
Reis, & C. M. Judd (Eds.), Handbook of research methods in social and
personality psychology (pp. 412–450). New York: Cambridge Univer-
Word, C. O., Zanna, M. P., & Cooper, J. (1974). The nonverbal mediation
of self-fulfilling prophecies in interracial interaction. Journal of Exper-
imental Social Psychology, 10, 109–120.
Zanna, M. P. (2004). The naı ¨ve epistemology of a working social psychol-
ogist (or the working epistemology of a naı ¨ve social psychologist): The
value of taking “temporary givens” seriously. Personality and Social
Psychology Review, 8, 210–218.
Zanna, M. P., & Cooper, J. (1974). Dissonance and the pill: An attribu-
tional approach to studying the arousal properties of dissonance. Journal
of Personality and Social Psychology, 29, 703–709.
Received June 23, 2004
Revision received July 1, 2005
Accepted July 21, 2005 ?
ESTABLISHING A CAUSAL CHAIN