Page 486 -

478 CHAPTER 15 Working with human subjects

15.3.5.2 Practical issues
Participants are crucial to our studies—without them, HCI research would be all but
impossible. We should make every effort to treat participants in a manner that reflects
this importance. Compensation for time and effort is certainly helpful, but research-
ers should also take concrete steps to make participation convenient and enjoyable.
Comfortable surroundings may put participants at ease. Ample opportunities for rest
or bathroom breaks should be provided, particularly for studies that involve longer
research sessions. Flexibility in scheduling and location can be particularly impor-
tant for some users: enrolling professionals in your study may require that you travel
to their workplace or allow for sessions outside of traditional working hours. If your
study is fun and convenient, participants may be more likely to help your recruiting
efforts by urging friends and colleagues to join in.
When working with human participants in any form of HCI research, you must
pay careful attention to your role as a researcher. Participants may be impressed or
intimidated by your presence, your use of language, your technical skills, the context
of the experiment, or any of a variety of related factors. This is particularly true for
observations and contextual inquiry, where you will spend a great deal of time in
close contact with one or more participants. Although you should make every reason-
able effort to help participants feel as at ease as possible, you should also be aware
that your presence may have an impact on observed performance. In some cases, par-
ticipants may exhibit the “demand characteristics” described above, trying to behave
in the manner that they think you are looking for.
Others have claimed that the mere act of participating in an experiment will influ-
ence user behavior, in the so-called “Hawthorne effect” (Macefield, 2007). Although
this effect has been the subject of significant debate among scientists (McCambridge
et al., 2014; Levitt and List, 2011), some suggested responses are appropriate.
Researchers should never give feedback regarding user performance during the
course of a study and experiments involving the comparison of multiple interfaces
should be controlled and “blind”—participants should not know if one of the alterna-
tives is favored by the researchers (Macefield, 2007).
More generally, these concerns about the influence of researchers on experimen-
tal results point towards a need to be modest about the results of our research. All
experiments have flaws and no single study establishes incontrovertible facts on its
own. When reporting results and drawing conclusions, we should avoid overstate-
ment, admit the flaws in our research, and point the way for future work that will
bring greater understanding.

15.3.6 ADDITIONAL CONCERNS
15.3.6.1 Potentially deceptive research
Does respect for persons always require complete disclosure regarding research
goals and design? Although it might seem as if withholding key details from research
participants might be somewhat less than fully honest, complete transparency might
not be appropriate in some cases, particularly if knowledge of the goals of the study