Anthony Galt Greenwald

Introduction

Anthony Galt Greenwald is an American social psychologist whose career runs across two of the largest research programs in modern social psychology: the study of how attitudes change, and the study of the attitudes people cannot or will not report. He is best known as one of the three co-creators of the Implicit Association Test (IAT), developed with Mahzarin Banaji and Brian Nosek and first published in 1998, which measures the strength of automatic associations between concepts by timing how quickly people sort them. The IAT became the most widely used instrument in the study of implicit bias, and the phrase "implicit bias" entered ordinary language largely through it. Greenwald spent two decades at Ohio State University and then, from 1986, at the University of Washington, where he is professor emeritus.

Early Life and Education

Greenwald was born in New York City, in 1939; a number of reference works give 1936 instead, and the two dates have never been fully reconciled in print. He received a B.A. from Yale University in 1959, then moved to Harvard University for graduate work, taking an M.A. in 1961 and completing a Ph.D. in psychology in 1963. His dissertation, supervised by the attitude researcher William J. McGuire, examined how commitment and free choice produce cognitive change — a question about belief change under constraint that stayed at the centre of his work for the next thirty years. From 1963 to 1965 he held a postdoctoral fellowship at the Educational Testing Service, where the exposure to psychometrics and large-scale testing shaped the methodological instincts that later distinguished his response-time work from the questionnaire tradition he grew up in.

Academic and Professional Career

Greenwald joined the psychology department at Ohio State University as an assistant professor in 1965, was promoted to associate professor in 1967 and to professor in 1971, and stayed until 1986. During those years he became one of the field's gatekeepers as well as one of its contributors: he was associate editor of the Journal of Personality and Social Psychology from 1972 to 1976 and its editor from 1977. In 1986 he moved to the University of Washington as professor of psychology, and it was there, in collaboration with Banaji, that his research turned decisively toward unconscious social cognition. He retired in 2020. He was associate editor of Experimental Psychology from 2001 to 2005, was elected to the American Academy of Arts and Sciences in 2007, and received the Donald T. Campbell Award from the Society for Personality and Social Psychology in 1994, a National Institute of Mental Health Research Scientist Award from 1998 to 2004, the Society of Experimental Social Psychology's Distinguished Scientist Award in 2006, the American Psychological Association's Award for Distinguished Scientific Contributions to Psychology in 2017, a Golden Goose Award in 2018 shared with Banaji and Nosek, and the BBVA Foundation Frontiers of Knowledge Award in Social Sciences in 2025, shared with Banaji and Richard Petty.

Major Contributions to Psychology

Attitude change and the cognitive response model

Long before the IAT, Greenwald was a major figure in attitude research. His 1968 chapter on cognitive learning and cognitive response to persuasion, written with Timothy Brock and Thomas Ostrom, reframed persuasion as something that happens inside the recipient: what determines attitude change is not the message but the thoughts the person generates in response to it. The argument that strong arguments persuade because they elicit favourable elaboration, and weak arguments fail because they elicit counterarguments, became standard in models of persuasion and advertising, and it supplied the "elaboration" step that the elaboration likelihood model of persuasion later built on.

The totalitarian ego

In 1980 Greenwald published "The totalitarian ego: Fabrication and revision of personal history" in American Psychologist, arguing that the self functions less like a historian than like a propagandist. People revise their memories of their own past attitudes and behaviour so that they appear consistent, benevolent, and in control. His formulation — that the past is remembered as though it were a drama in which the self was the leading player — connected the study of memory to the study of self-presentation, and anticipated later work on motivated reasoning and self-serving bias.

In 1995 Greenwald and Banaji published "Implicit social cognition: Attitudes, self-esteem, and stereotypes" in Psychological Review, the paper that gave the field its name. They argued that social judgments are frequently driven by traces of past experience that operate without awareness, intention, or control, and that these implicit attitudes can diverge from what a person reports. The paper was a synthesis of scattered findings on subliminal priming, mere exposure, and automatic evaluation rather than a new experiment, and its influence came from giving a diffuse literature a single label and a research agenda.

The Implicit Association Test

The measurement problem followed directly. If implicit attitudes are introspectively inaccessible, they cannot be measured by asking. In 1998 Greenwald, Debbie McGhee, and Jordan Schwartz published the IAT in JPSP: participants sort words and images into paired categories as fast as possible, and the difference in response latency between two pairing conditions estimates the relative strength of the underlying association. Because it is a performance measure rather than a self-report, it is difficult to fake deliberately. Greenwald, Nosek, and Banaji published an improved scoring algorithm in 2003, and Greenwald and his collaborators launched Project Implicit, a public website that let anyone take the tests and simultaneously donate their data to research. Between October 1998 and October 2006 more than 4.5 million tests were completed there, an unprecedented dataset for research on intergroup attitudes.

Major Publications

Greenwald, A. G., & Banaji, M. R. (1995). Implicit social cognition: Attitudes, self-esteem, and stereotypes. Psychological Review, 102(1), 4–27.
Greenwald, A. G., McGhee, D. E., & Schwartz, J. L. K. (1998). Measuring individual differences in implicit cognition: The implicit association test. Journal of Personality and Social Psychology, 74(6), 1464–1480.
Greenwald, A. G. (1980). The totalitarian ego: Fabrication and revision of personal history. American Psychologist, 35(7), 603–618.
Greenwald, A. G., Nosek, B. A., & Banaji, M. R. (2003). Understanding and using the Implicit Association Test: I. An improved scoring algorithm. Journal of Personality and Social Psychology, 85(2), 197–216.
Greenwald, A. G., Poehlman, T. A., Uhlmann, E. L., & Banaji, M. R. (2009). Understanding and using the Implicit Association Test: III. Meta-analysis of predictive validity. Journal of Personality and Social Psychology, 97(1), 17–41.
Greenwald, A. G., & Banaji, M. R. (2013). Blindspot: Hidden Biases of Good People. Delacorte Press.
Greenwald, A. G., Banaji, M. R., & Nosek, B. A. (2015). Statistically small effects of the Implicit Association Test can have societally large effects. Journal of Personality and Social Psychology, 108(4), 553–561.

Impact and Legacy

The IAT did something few psychological instruments have done: it left the laboratory. Corporations adopted it for diversity training, courts and commentators cited it in discrimination debates, and "take the implicit bias test" became a common piece of advice. Greenwald himself repeatedly warned against the use the test was being put to, and Nosek, Greenwald, and Banaji stated in their own 2007 review that the IAT is not designed to diagnose or label an individual.

The scientific criticism has been sustained and is worth stating precisely. The central question is whether IAT scores predict discriminatory behaviour. Greenwald's own 2009 meta-analysis, covering 122 reports and 14,900 participants, reported an average correlation of r = .274 between IAT scores and behavioural, judgment, and physiological criteria (r = .236 for Black/White interracial behaviour); in the 156 samples that also included self-report measures, the explicit measures predicted at r = .361, that is, better than the IAT. An independent meta-analysis by Oswald, Mitchell, Blanton, Jaccard, and Tetlock in 2013, restricted to race and ethnicity outcomes and recoding which criteria actually counted as discrimination, put the figure at about .14 and concluded that the IAT provides little insight into who will discriminate against whom and no more insight than explicit measures. When Carlsson and Agerström re-analysed the same corpus in 2016 using a stricter definition of discrimination, the effect was close to zero and highly inconsistent. Later large syntheses have produced estimates as low as .097. The test's test-retest reliability, around .5, is well below that of good self-report scales, which matters most in exactly the individual-level applications it is popularly used for.

Greenwald and his co-authors replied that statistically small effects can have societally large effects, because small biases cumulate across repeated encounters and across many decision-makers. Critics such as Ulrich Schimmack have countered that this reply treats a correlation as a causal path: IAT scores do not cause behaviour, and if part of the IAT's variance is not valid, then a correlation of .1 is a lower bound on the effect of prejudice rather than the effect of the test. A second line of attack concerns intervention: Forscher and colleagues' 2019 review of hundreds of studies found that implicit measures can be shifted in the laboratory but that the shifts rarely translated into lasting behavioural change, which undermines the assumption that lowering IAT scores is a route to lowering discrimination. A third response, from within the implicit-bias tradition itself, has been to reinterpret what the test measures. Payne, Vuletich, and colleagues have argued that IAT variation may index the bias of situations and environments a person has absorbed rather than a stable property of the person, which would preserve the test's value for studying contexts while conceding the critique of it as an individual diagnostic.

Whatever the resolution, the research programme Greenwald started changed the terms of the discussion. Discrimination could no longer be explained only by people who knowingly dislike outgroups, and a generation of measurement work — scoring algorithms, reliability estimation, pre-registered replication — was carried out in direct response to his instrument.

Learn More

Leon Festinger — whose cognitive dissonance work set the agenda for decades of social cognition research.
Jennifer Eberhardt — whose research on racial bias and policing extends the implicit-association line.
Daniel Kahneman — whose heuristics programme reshaped the same decades of judgement research.
Social Cognition — the field built around how people encode and interpret social information.

References

Greenwald, A. G., McGhee, D. E., & Schwartz, J. L. K. (1998). Measuring individual differences in implicit cognition: The implicit association test. Journal of Personality and Social Psychology, 74(6), 1464-1480.