HOW TO REVIEW THE LITERATURE AND CONDUCT ETHICAL STUDIESPhysical harm is rare, even in biomedical research,
in which the intervention is much greater. Specific
types of harm are more likely in different types of
research (e.g., in experimental versus field research).
Researchers must be aware of all types of harm and
work to minimize them at all times.^12
Physical Harm. A core ethical principle is that
researchers should never cause physical harm to
participants. This means we must anticipate risks
before beginning research, including basic safety
concerns (safe buildings, furniture, and equipment).
We screen out high-risk subjects (those with heart
conditions, mental illness, or seizure disorders) if
the study involves stress, and anticipate the danger
of injury and even physical attacks on research par-
ticipants or assistants. We accept moral and legal
responsibility for any injury that occurs as a result
of research participation. This means that we must
immediately terminate a study if we cannot guar-
antee the physical safety of particpants (see the
Zimbardo study in Example Box 4, Three Cases of
Ethical Controversy).
Psychological Abuse, Stress, or Loss of Self-Esteem.
Although the risk of physical harm is rare, social
researchers may place people in stressful, embar-
rassing, anxiety-producing, or unpleasant situa-
tions. To learn about how people respond in
real-life, high anxiety-producing situations, social
researchers have placed research participants in
realistic situations of psychological discomfort or
stress.The ethics of the famous Milgram obedience
study is still debated (see Example Box 4). Some
say that the precautions taken by Milgram and the
knowledge gained outweighed the stress and poten-
tial psychological harm that research participants
experienced. Others believe that the extreme stress
and the risk of permanent harm were too great.
Some researchers have created high levels of
anxiety or discomfort by exposing participants
to gruesome photos, falsely telling male students
that they have strongly feminine personality traits,
falsely telling students that they have failed, creat-
ing a situation of high fear (e.g., smoke entering a
room in which the door is locked), asking par-
ticipants to harm others, placing people in a situa-
tion in which they face social pressure to deny their
convictions, and having participants lie, cheat, or
steal.^13 Researchers who study helping behavior
may place participants in emergency situations to
see whether they will lend assistance to “victims.”
For example, Piliavin and associates (1969) studied
helping behavior in subways by faking someone’s
collapse onto the floor. In the field experiment, the
riders in the subway car were unaware of the exper-
iment and did not volunteer to participate in it.
A sensitive researcher is also aware of harm to
a person’s self-esteem. For example, Walster (1965)
wanted to see whether feelings of female self-worth
affected romantic liking. She gave undergraduate
women personality tests followed by phony feed-
back. She told some that they lacked imagination
and creativity. Next, a handsome male graduate stu-
dent who pretended to be another research partici-
pant struck up a conversation with the women. He
acted very interested in one woman and asked her
out for a dinner date. Walster wanted to measure the
woman’s romantic attraction to the male. After the
experiment, the woman learned that there was no
date and the man was just working in an experiment
and was not really interested in her. Although the
participants were debriefed, they suffered a loss of
self-esteem and possible psychological harm.^14
Only experienced researchers who take pre-
cautions before inducing anxiety or discomfort
should consider conducting studies that induce
stress or anxiety. They should consult with others
who have conducted similar studies and mental
health professionals when planning the study,
screen out high-risk populations (e.g., people with
emotional problems or a weak heart), and arrange
for emergency interventions or termination of the
research if dangerous situations arise. Researchers
should always obtain informed consent (to be dis-
cussed) before the research and debrief the subjects
immediately afterward.
A core ethical principle is that researchers
should never create unnecessarystress in partici-
pants. Unnecessarymeans beyond the minimal
amount required to create the desired effect, or stress
without a direct, legitimate research purpose. Know-
ing the minimal amount comes with experience. It
is better to begin with too little stress, risking find-
ing no effect than to create too much. If the level of