Using the Online Psychology Laboratory (OPL) to illustrate the Concept of Implicit
Attitudes by Pamela Marek, Kennesaw State University
This third article in the OPL (http://opl.apa.org/) series focuses on involving
students in a procedure for assessing attitudes that was designed by Greenwald, McGhee,
and Schwartz (1998) to overcome the bias present in self-report techniques. Because it
involves associations that are activated outside of our conscious awareness, the Implicit
Association Test is an appropriate activity to introduce when discussing consciousness.
However, the demonstration is directly linked to prejudice; thus, introductory psychology
textbooks typically include information about such implicit measurement in the social
psychology chapter.
Attitudes, evaluative judgments, are an important focus of social psychology
coverage in most introductory textbooks. Frequently, attitudes are examined in an ABC
framework, involving affective, behavioral, and cognitive correlates and consequences
(Olson & Zanna, 1993). Insofar as the attitude of prejudice, an evaluative judgment of a
person or object based solely on group membership, the affective manifestation of
negative affect might include hostility, anxiety, or even fear. The cognitive correlate
might be a negative stereotype, with people believing that the stigmatized group is lazy,
aggressive, or unmotivated. The behavioral manifestation might involve a predisposition
to discriminate against, or behave negatively toward, the stigmatized group. Knowledge
of the distinction between prejudice and discrimination may help students better
understand the meaning of their IAT results.
Students who are typically willing to express their attitudes on a wide variety of
topics may not readily realize that at times they may be unaware of their attitudes. In fact,
explicit attitudes, those that students are able and willing to express, may conflict with
implicit attitudes, those that students are unaware of or are unwilling to express.
Prejudice is a prototypical area in which implicit and explicit judgments may conflict.
Explicit measurement of prejudice, whether related to age, sex, ethnicity, or sexuality, is
clearly prone to social desirability effects. Few people, with the possible exception of
members of hate groups, would publicly admit to holding prejudiced attitudes against or
negative stereotypes of older people, women, minority groups, or homosexuals.
Instructors might introduce students to early techniques for circumventing social
desirability by explaining an experiment involving a bogus pipeline, perhaps linking an
experimental description to coverage of research methods. For example, Sigall and Page
(1971) used a bogus pipeline to assess stereotypes, randomly assigning Caucasian men to
one of two groups. One group rated “Americans” and “Negroes” on a series of traits
using a standard pencil and paper test, whereas the other group were led to believe that a
wheel that they turned to indicate their ratings could detect if they were untruthful. This
bogus pipeline group held more negative stereotypes of Negroes and more positive
stereotypes of Americans than did the pencil and paper control.
More recently, Greenwald et al. (1998) developed an implicit association test that
has been used to measure implicit attitudes, defined as “introspectively unidentified (or
inaccurately identified) traces of past experience that mediate favorable or unfavorable
feeling, thought or action toward social objects” (Greenwald & Benaji, p. 8). The test is
based on the principle that people will respond more readily to pairs of concepts that they
associate with each other, even if they are not consciously aware of the association. The
on-line psychology lab includes such a test to measure implicit or automatic prejudice
toward African Americans and Euro Americans. When introducing this activity to
students, instructors might position it as a demonstration of an implicit measure of
attitudes, rather than as a “test”. To some extent, this might assuage potential student
discomfort in participating in what they might view as a sensitive activity.
Prior to distributing the assignment to students, instructors are encouraged to visit
the OPL website to obtain a class ID. This will enable instructors (or students) to
download class data for further examination after the activity is completed. When
students begin the demonstration, they are presented with two optional questions about
explicit attitudes (preference for Whites vs. Blacks using a Likert-type scale and a
“thermometer” measure of warmth-coldness toward Blacks and Whites). Next, to prepare
for the IAT, students place one finger on each of two keys on opposite sides of the
keyboard. For responses to be meaningful, students are instructed to proceed quickly. The
actual demonstration contains five blocks of trials, summarized in the table below.
Block
1
2
3
4
5
Task
Initial target
concept
discrimination
Associated
attribute
discrimination
Initial
combined task
(Stereotypecongruent)
Reversed
target concept
discrimination
Reversed
combined task
(Stereotype
noncongruent)
Stimulus
Key Stimulus
African-American Left European
face
American face
Key # of trials
Right 10
Bad words (e.g.
“agony”)
Right 10
Left Good words
(e.g. “joy”)
African-American Left Europeanor bad words
American of
good words
Right
EuropeanAmerican face
Left AfricanAmerican face
Right
EuropeanAmerican face or
bad words
Left AfricanAmerican face
or good words
Right
For analysis, the key blocks of interest are the two combined tasks (Blocks 3 and
5). The IAT is scored by calculating the difference in average reaction time for Block 3
(the stereotype congruent condition) and the average reaction time for Block 5 (the
stereotype non-congruent condition). A positive score is interpreting as suggesting an
automatic association between African-Americans and negative terms, or as a preference
for European-Americans. At the end of the demonstration, students are shown their IAT
score and given the opportunity to save their data.
Instructors are strongly encouraged to allot class time to discuss the results of the
demonstration and to respond to potential student skepticism about the meaningfulness of
this research technique. First, it is important to reassure students that a positive score does
not mean that they display either prejudice or discrimination. Rather, Devine (1989) has
demonstrated that whereas most people are aware of negative stereotypes about AfricanAmericans that may be activated automatically, people’s personal beliefs may conflict
with the negative stereotype. Given time to consciously access their personal beliefs, lowprejudiced people may invest cognitive effort to inhibit the negative stereotype, thereby
avoiding unpleasant affect or discriminatory behavior. In contrast, high-prejudiced people
would not attempt to consciously inhibit prejudice or discrimination because their
personal beliefs coincide with the negative stereotype. On the IAT, the rapid response
requirement is purposely designed to eliminate the possibility of conscious inhibition of
stereotypes, thereby measuring only the automatic component of the attitude of prejudice.
Second, students may complain that, despite the Block 4 practice at reversed
target concept discrimination, they had difficulty making accurate rapid classifications in
the combined Block 5 task because the designated keys for European American and
African American faces had been reversed from Block 3. Instructors might acknowledge
that this is, indeed, a valid point, consistent with research findings that have revealed
small order effects, with faster response latencies for whichever combined task is
presented first. Instructors might take this opportunity to make a connection between
proactive interference in memory and the manner in which interference may influence
performance on the second target concept discrimination task. Instructors might also
mention that some on-line versions of the IAT (e.g., Project Implicit at
https://implicit.harvard.edu/implicit/ ) randomly select the ordering of the combined tasks
to reduce the influence of order effects on aggregate scores for groups of participants.
Another task alteration that reduces the order effect involves increasing the number of
trials in Block 4, the reversed target-concept discrimination task (Nosek, Greenwald, &
Banaji, 2005).
For further discussion, instructors (or students) may opt to download the class
data as an Excel file, in which the mean response latencies (in seconds?) for the
stereotype congruent and stereotype noncongruent conditions may be computed and
compared. It is expected that times will be faster for the stereotype congruent condition,
suggesting the presence of an automatic association. Supporting this hypothesis, in a
fascinating fMRI study in which response times on the IAT were linked to activation of
the amygdala, Phelps et al. (2000, p. 731) included the following graph to illustrate the
IAT results (in milliseconds rather than seconds). Does class data resemble this pattern?
Note that Phelps et al. might serve as a link to coverage in biological psychology chapter
relating the amygdala to emotional responses.
RT (ms.)
1000
750
500
250
Stereotype
noncongruent
Stereotype
congruent
Condition
An excellent follow up activity to the IAT involves directing students to the
Southern Poverty Law Center’s Tolerance website,
http://www.tolerance.org/hidden_bias/). By clicking on the link “Track U. S. hate
groups” in the Tolerance Watch box on the left, students can engage in an eye-opening
experience as they determine the variety of currently active hate groups in their home
state and the beliefs of each of these groups. On a more positive note, clicking on links in
the Do Something box can lead to the discovery of “10 ways to fight hate” or “10 ways to
fight hate on campus.” Students might discuss the proposed solutions in class or by
writing a brief paper relating the material to their textbook reading. Clearly, this site
provides a wide range of opportunities for creative class-based on community-based
assignments.
Overall, the IAT activity provides students with an introduction to a cutting-edge,
albeit controversial, research technique that has been used to study bias in a wide range of
personally relevant areas. It is a stimulating tool with applications that extend well
beyond introductory psychology to more advanced courses (e.g., social psychology,
cognitive psychology, and research methods.) An abundance of material concerning IAT
development, validity, applications, and controversies is available via links on Dr.
Anthony Greenwald’s homepage (http://faculty.washington.edu/agg/iat_materials.htm).
References
Devine, P. G. (1989). Stereotypes and prejudice: Their automatic and controlled
components. Journal of Personality and Social Psychology, 56, 5-19.
Devine, P. G., Plant, E. A., Amodio, D. M., Harmon-Jones, E., & Vance, S. L. (2002).
The regulation of explicit and implicit race bias: The role of motivations to
respond without prejudice. Journal of Personality and Social Psychology, 82,
835-848.
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, 1464-1480.
Nosek, B. A., Greenwald, A. G., & Banaji, M. R. (2005). Understanding and using the
Implicit Association Test: II. Method variables and construct validity. Personality
and Social Psychology Bulletin, 31, 166-180.
Olson, J. M., & Zanna, M. (1993). Attitudes and attitude change. Annual Review of
Psychology, 44, 117-144.
Phelps, O’Connor, Cunningham, Funayama, Gatenby, Gore, & Banaji (2000).
Performance on indirect measures of race evaluation predicts amygdale
activation. Journal of Cognitive Neuroscience, 12, 729-738.
Sigall, H., & Page, R. (1971). Current stereotypes: A little fading, a little faking. Journal
of Personality and Social Psychology, 18, 247-255.