Page 484 -

476 CHAPTER 15 Working with human subjects

for research projects involving genetic studies of biological samples have led to
the development of simplified consent forms (Beskow et al., 2010) and studies of
comprehensible phrasing for relevant content (Beskow et al., 2015). HCI research-
ers tackling deceptive studies have struggled with similar challenges (see Section
15.3.6.1). The growth of ubiquitous computing research, with many studies of con-
textualized computer and device use or interaction in use in homes, schools, and
other public places has led some to suggest that consent should be more nuanced
and contextualized than a simple blanket agreement to the terms of a research study
(Luger and Rodden, 2013a,b).

15.3.5 RESPECTING PARTICIPANTS

The Belmont report describes respect for persons in terms of ensuring that partici-
pants are able to make independent and informed decisions about their involvement
in research studies. Although this is perhaps most directly applied through the in-
formed consent process described above, truly respecting participants requires con-
sideration of their needs, concerns, and values throughout all aspects of designing the
study, conducting the research, and publishing the results.

15.3.5.1 Study design
The experiments conducted by Milgram and Zimbardo arguably arose at least in part
because their interest in their research question overshadowed concerns that they
may have had about the impact that the research would have on the participants.
Although these studies would clearly be considered unethical by current standards,
controversies regarding the impact of research on participants still rage. Examination
of these debates can shed some light on the challenges raised by some HCI research
projects.
A newsworthy study published in July 2014 by researchers from Cornell
University and Facebook (Kramer et al., 2014) provides a textbook example of how
these concerns might arise in modern HCI research. In order to understand the im-
pact of “emotional contagion”, these researchers worked with Facebook to manipu-
late the presentation of items on users’ news feeds. Over the course of 1 week in
Jan. 2012, researchers adjusted news feeds, decreasing either the amount of positive
or negative emotional content, as determined by the inclusion of words previously
shown to be correlated with measures of well-being. Examining news feeds for a
large (almost 700,000) group of Facebook users, they found that reductions in posi-
tive posts appearing on a users’ feed were associated with reductions in positive con-
tent in that individuals postings, with comparable effects for reductions in negative
content (Kramer et al., 2014).
The Facebook paper raised a firestorm of research ethics controversy immediately
upon completion. An “Editorial Expression of Concern” (Verma, 2014) published
alongside the paper noted the major concern: the research was conducted without ex-
plicit informed consent from participants. The study was consistent with Facebook’s
data use policies, which describe creation of an account as implicit agreement to