Available online at www.sciencedirect.com
International Journal of Drug Policy 19 (2008) 246–247
Response
Responses on a cautionary tale concerning the ethics of using
respondent-driven sampling to study injection drug users
Greg Scott
*
DePaul University, Chicago, IL, United States
Received 25 March 2008; accepted 25 March 2008
First I want to thank my colleagues for responding so pas-
sionately to my recently published article (Broadhead, 2008;
Lansky & Mastro, 2008; Ouellet, 2008; Prachand & Benbow,
2008; Scott, 2008). The Journal Editors have presented me
with the opportunity to reply. A close reading of the article
should lead to the conclusion that I am not attacking respon-
dent driven sampling (RDS) or people who use it, much less
the people who “invented” it. As early as the abstract I clearly
state my belief that RDS remains the best available means
for sampling hidden populations. It does not follow, how-
ever, that RDS should be sheltered from critical scrutiny. My
goal has been to examine RDS’s inner workings with an eye
toward improving the method’s capacity for accruing repre-
sentative samples while simultaneously maximizing human
subject protections.
Ouellet
First, sero-mixing of IDU networks occurs constantly. I
report that it may have occurred in response to RDS pro-
pellants. I clearly indicate that I have no sero-data proving
that it did, in fact, occur as a result of RDS. Second, my
article addresses some of the harms associated with this sam-
pling method and some of the potential harms associated with
RDS. Third, Ouellet (2008) insinuates that “Stan” probably
could not have recalled having given “Patty” a coupon sev-
eral weeks earlier without my having asked about it. The
claim is problematic because (1) this incident occurred before
Stan and Patty knew anything about my involvement in the
RDS evaluation (therefore an “observer effect” could not
be an accomplice), and (2) Ouellet states that street-level
*
Tel.: +1 773 882 0937; fax: +1 773 882 0937.
E-mail address: gscott@depaul.edu.
addicts “live in the now” and ostensibly do not or cannot
cull from longer term memory the debts owed to them. Street
ethnographers know that “street addicts” survive in part by
maintaining a robust capacity for keeping a detailed mental
balance sheet of debits and credits.
Fourth, Ouellet claims that National HIV Behaviour
Surveillance (NHBS) staff members know “Wilma” and that
she “was given to considerable exaggeration.” This statement
is ludicrous. Wilma is a pseudonym for a woman whose real
first name (or any other identifying traits) I never recorded,
as per the protocol approved by DePaul University’s Insti-
tutional Review Board. Fifth, Ouellet finds it “difficult to
imagine that all 20 shooting gallery operators engaged in
virtually the same hustle.” The emulation of others’ hustles
happens constantly on the street. Those of us who con-
duct extensive street ethnography among addicts know how
rapidly such enterprising schemes can spread through affil-
iated and disparate social networks and across geographic
areas.
Broadhead
Broadhead (2008) concludes that my work is “decontex-
tualized” and that I am “oblivious, reckless and self-serving.”
Rather than entertaining a critique concerned principally
with advancing a moralistic personal condemnation, I simply
want to explain that this critic – like the others – misses the
point of my article. First, I am not “bashing” RDS but rather
exploring the manifold ways that RDS allows participants
to secure more relative control over the economic system
embedded in the sampling design. Such “empowerment”
is what makes RDS the best means for accessing hidden
populations. But it also engenders opportunities for various
forms of exploitation.
0955-3959/$ – see front matter © 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.drugpo.2008.03.008