Page 472 -

464 CHAPTER 15 Working with human subjects

make them well equipped to understand the reasons for conducting the experiment
and the potential costs and benefits involved in participation in a study. Potential
participants may lack some or all of this relevant background.
Research studies should be designed to protect participants. Informed con-
sent—the notion that research participants should be provided with the infor-
mation needed to make a meaningful decision as to whether or not they will
participate—is the cornerstone of this protection. Academic and industrial or-
ganizations that conduct research with human subjects generally rely on institu-
tional review boards (IRBs) to review proposed research for any possible risks
and to guarantee that appropriate procedures for informed consent are being
followed.

15.3.1 RISKS AND CONCERNS OF RESEARCH PARTICIPANTS
Participation in a research study involves multiple agreements between the partici-
pant and the researcher. The participant agrees to perform certain tasks as needed by
the experiment and the experimenter frequently agrees to provide some incentive or
compensation to the participant. Perhaps more importantly, experimenters agree to
conduct responsible research that protects participants' rights, health, privacy, and
safety.
Risks to participants are often most pronounced in medical research, where in-
vestigation of new drugs, devices, and procedures can lead to health risks, particu-
larly when things don't work as intended (or hoped). However, physical harm is
not necessarily the only relevant concern. Famous psychology experiments have
shown how research that places people in uncomfortable situations can cause sig-
nificant emotional distress (see the Milgram's Experiment and Stanford Prison
Experiment sidebars). Although some HCI experiments might raise these con-
cerns, most of the studies in our field are low risk. Some studies may lead to fa-
tigue (from mouse movements) or eye strain, but these risks are minor. Regardless
of the level of risk involved, researchers must treat human participants with dig-
nity and respect.

MILGRAM'S EXPERIMENT

Perhaps the most famous example of deception in psychology research, Stanley
Milgram's obedience experiment illustrates one possible extreme associated
with research on human subjects.
In this study, subjects were told that they were participating in a study of the
effect of punishment on learning. They were asked to administer tests to another
subject—a “learner”—who would have to identify a word that had previously
been associated with a stimulus word. Subjects were told that they had to