IP: Ethics Questions for CMU, Rimm, and Sirbu part 2 of 2 (limits

[David Farber <farber () central cis upenn edu>]

by voice and e-mail.  Robert Thomas, sysop of Amateur Action BBS,
currently serving a sentence in a Federal penitentiary for making
available adult material on his system that, while not illegal in his
own state of California was illegal in Tennessee, was one of Rimm's
research subjects.  Rimm's account of events does not correspond with
e-mail corespondence between Rimm and Thomas provided by Thomas's
wife (CuD 7.59).


*QUESTION: Did Rimm in fact communicate  with all the sysops as he
claimed?


*QUESTION: Did Rimm's communication with the sysops indicate the
kind of ethical impropriety that the released e-mail between he
and Robert Thomas suggests?


*QUESTION: Why doesn't the Committee of Investigation talk with
Robert Thomas in order to ascertain how data from AA BBS were
acquired?


Early attempts to obtain a copy of the study from Rimm or the
Georgetown Law Journal were barred by claims that the study was
embargoed. In a December 18, 1994 post to Mike Godwin, Rimm claimed
that the GLJ embargoed the study. In posts on The Well, a popular
public access system in California, Time's Philip DeWitt claimed
that he had an exclusive with Rimm. Kathy Ruemler, current
editor-in-chief of the GLJ, wrote in a public Usenet post on
September 6:


     V.  RUMOR:  TIME was restricted from having the study
     independently reviewed by an agreement with Law Journal.
          FACT:  The Law Journal had no such agreement with TIME.
     Isn't TIME the one who referred to it as an exclusive?


*QUESTION: Somebody seems to be lying. Who?


An argument could be made that Rimm's advisor, Marvin Sirbu, and
not Rimm, ought bear responsibility for improprieties in the
study.  After all, an advisor is ultimately responsible for
assuring that proper ethical and methodological procedures are
followed.  In this case, such a judgment might be premature.


In 1984, Rimm was involved in a study of high school gambling with
Henry Lesieur, then a sociology professor at Johns Hopkins, and Bob
Klein, reportedly a high school counsellor.  Lesieur apparently was
drawn to Rimm by his infiltration of a casino at 16:  "That just
intrigued me," he said.


The study involved giving students in five high schools
questionnaires. Rimm distributed them in one high school, Lesieur
said.  Was Rimm sufficiently apprised of ethical issues a decade ago
such that he should be aware of appropriate behavior now?  Lesieur
could not say, but he observed:


     We had meetings and we went through the (ethical) protocols.
                             .........
     People didn't have to respond, it was totally anonymous.
     It (the study) went through human subjects, he was part
     of the process, and he followed the protocols.


Although Rimm was a third author on a paper, Lesieur indicated that
Rimm, in fact, did no writing.


Now, it's unreasonable to expect a college undergraduate to fully
understand the nuances of research ethics, let alone recall them a
decade later. But, it's not unreasonable to expect that, given this
apparent background in research, Rimm would not be aware that there are
ethical protocols. Therefore, those who see Rimm as a "victim" of
inadequate supervision have a weak case: Rimm was in a position to
know that there are guidelines for protecting human subjects and that
his own methodological descriptions indicates that he violated them.
However, this still leaves several questions that Professor Sirbu
might clear up.


                       QUESTIONS FOR SIRBU


Although Sirbu was quoted in a July New York Times story as saying
that he never saw the final article that was submitted to GLJ, and
that it was not the report he would have written, it is clear that
he was closely involved with Rimm throughout the study.  At issue
here isn't the final article, but how Rimm could continue to collect
data in ways that raise serious questions about why the advisor,
who was professedly close to the study, did not engage in
corrective intervention.


In November correspondence with Mike Godwin, Sirbu claimed that he
had no problem with the methodology and would be glad to discuss it.
But, Sirbu seemed well-aware of the study long before that,
as a memo to several CMU faculty and administrators indicates:


     Date: Tue, 27 Sep 1994 23:11:43 -0400 (EDT)
     From: Marvin Sirbu <ms6b+ () andrew cmu edu>
     To: Erwin Steinberg <es2t+ () andrew cmu edu>,
         Michael Caldwell Murphy <mm1v+ () andrew cmu edu>,
         Don Hale <dh0c+ () andrew cmu edu>
     Subject: Martin Rimm's research
     Cc: Jessie Barbour Ramey <jr3l+ () andrew cmu edu>,
         Martin Rimm <mr6e+ () andrew cmu edu>


     Gentlemen,


     I understand that you have been in touch with Martin Rimm
     regarding his research.  Since Martin is currently working
     on this for credit under me, I have asked him if he would
     permit me to be included in any meetings that you may
     arrange.


     I have been meeting regularly with Martin since last Spring,
     and believe that he is nearing completion on a
     ground-breaking study that makes an important scholarly
     contribution.  He has developed some very interesting
     methodological approaches, and has amassed a remarkable
     database of information on his chosen subject matter.  As
     Martin and I have discussed, there is still much to do in
     interpreting the data.


     The bulk of his data collection focuses on privately
     operated Adult Bulletin Board Systems (BBS's) offering
     sexually oriented imagery.  He has also examined partial
     data on the availability and consumption of such imagery at
     CMU from the Internet, although this data is not central to
     his work.


     We have had numerous discussions as to the most appropriate
     venue for publishing this work since it may appeal to groups
     as diverse as those concerned with telecommunications
     policy, law, mass communications, marketing science, or
     sociology of sexual deviance.  We have also been discussing
     potential sources of external research support.  Our most
     recent thinking has been to produce a Working
     Paper/Technical Report that could be disseminated from CMU
     pending determination of the most appropriate avenue for
     formal publication.


     Because of the subject matter, this research could provide
     fodder for everyone from the Kinsey Institute to Jerry
     Falwell to Andrea Dworkin, as have previous scholarly
     studies in this field.


     I might not have chosen myself to raise these issues via a
     message directly to the President, but sooner or later this
     study will come out and I suspect there will be significant
     interest among the press.  It is certainly appropriate that
     CMU be prepared.  Martin and I both concur that the way the
     research is publicized should be handled with great care,
     but I know that he is anxious, after working on this for
     more than a year, to get something out before he starts
     applying to graduate schools this fall.


Among other things, Sirbu reveals his knowledge of Rimm's access to
"availability and consumption" of the Usenet readership habits of
CMU system users.  It is well-established that users have a reasonable
expectation of privacy. Sometimes, fulfilling administrative duties
requires system administrators to monitor use, files, or other
material that a user intends to be private. However, sometimes such
monitoring raises questions. In the Rimm study, for example, users'
Usenet newsgroup configuration files were systematically tabulated.
Although reports differ on whether systems engineers or third parties
monitored the files, it is clear that, according to the Rimm study,
data on individuals were collected and compared, and the aggregate
results then made public.  It remains unclear whether, despite the
serious appearance of impropriety, any breaches occurred. However, the
question is of sufficient import to be addressed:


*QUESTION: Did the acquisition of individual user information as
described in Rimm's methodology, actually occur? If so, is such
acquisition consistent with the ethical guidelines on human subjects?


If Sirbu were as close to Rimm's study as his public pronouncements
and private correspondence indicate, he surely would have, or should
have, known of the practices Rimm employed.


*QUESTION: "What did Sirbu know and when did he know it?"


As Sirbu should know, "research team" has a special connotation among
scholars. A research team is not a casual circle of people who may
occasionally interact.  Sirbu's professed close relationship with Rimm
and involvement in Rimm's research would give him knowledge of
whether a "research team," as the term is conventionally employed by
reputable scholars, did in fact exist. Sirbu's claim (above) that
some high-level faculty "collaborated" with Rimm adds credence to,
and perpetuates the image of, an established group of professionals
well-integrated into a research project directed by Rimm as
"principal investigator."


Given the fact the some "team" members were unaware that they were
team members or have denied that they were members at all, one cannot
help but suspect that the public is being deceived into believing
that the study is more credentialed than it actually is.


*QUESTION: Can Sirbu explain precisely what the "research team"
members did to justify the label?


*QUESTION: If there was, in fact, no "research team" in the
conventional use of the term, why did Sirbu allow the fiction to
persist?


Perhaps the most important question Sirbu could address is the
attempt to acquire funding for Rimm's project. It appears that
Sirbu's attachment to the study included attempts to ride the funding
gravy train by cashing in on Rimm's methodology.


In November, Sirbu approached EFF's Mike Godwin to solicit EFF
support for the project:


      Date: Mon,  7 Nov 1994 22:05:16 -0500 (EST) From: Marvin
      Sirbu <ms6b+ () andrew cmu edu> To: mnemonc () eff org Subject:
      Your visit to CMU


      As you may have gleaned from reading about the events at
      CMU, I have been working with Martin Rimm on a study of
      the availability and consumption of sexually explicit
      imagery on Adult BBS systems and, to a lesser extent, on
      Usenet.  Xxxx Yyyy suggested that EFF might be interested
      in the work we've been doing.  Among other things, we
      have data which could be analyzed to show the geographic
      distribution of consumers of adult BBS systems.  Such
      data might be useful in countering or confirming
      assertions that "community standards" in places like
      Memphis are different from other regions of the country.


      I'll give you a call when we are both back in our
      respective cities.


      Marvin Sirbu


Neither EFF's mission nor resources allow for such support, and the
solicitation was rejected. But, federal funding remained a
possibility.


David Banks, a CMU statistics professor, provided some
technical guidance for Rimm. According to Banks, in early November,
1994, he, Sirbu, and Rimm met to discuss what might be done with the
paper. Rimm, reports Banks, needed money, because he sunk some of
his own funds into the project. According to Banks:


     Also, Martin was aware that the Department of Justice
     had cut down AA BBS and seized their log files. And that
     info should contain names, log files, and it seemed
     reasonable to suspect that DOJ would have that set
     analyzed and that they would pay money for it.


     The grant attempted to link our interests with DoJ
     prosecutorial interests.


According to Banks, the proposal had four research goals:


1) A summary of the statistics of "pornography" traffic that
   would identify the proportion of BBSes with a high
   percentage of material that might be worth prosecuting;


2) Consumption and usage trends over time: If pornography or
   pedophilia increases, then it would indicate that the BBS
   is trying to cultivate that market;


3) Information on individual downloads and covariance of
   user preferences that would correlate which types of
   files are most-likely to be associated other downloaded files;


4) "Placing it in the space of adult bulletin boards; adult
    BBSes have different personalities, characteristics, and
    specialties...who is the worst offender on pedophilia?"


This, Banks explained, would allow DoJ to distribute its
prosecutorial resources more effectively.


Rimm was not listed as a co-principal investigator on the second
round of grant submission, Banks explained, because DoJ would run
the grant through CMU, which would be more difficult if Rimm were
not a student.  However, Rimm was written in as a consultant,
according to Banks.


Banks said that he often expressed his concern with ethical issues,
both orally and in writing, and in July, 1995, he withdrew from the
project because of these concerns.


Why is the grant significant?


The grant application raises serious ethical questions for Sirbu and
Rimm. One fundamental canon of accepted ethical procedures is that
researchers do nothing to put their subjects at risk (see CuD 7.58).
Yet, that is precisely what this grant application would do.


In his methodology, Rimm explains that he selected BBSes that were
either the largest and most active "pornography" distributors, or
that appeared to be aggressively moving into the "pornography" market
(GLJ, 1995:  1876-77). If Banks's summary of the DoJ grant proposal is
accurate, these BBSes are precisely those that the grant was designed
to help prosecute, because they constitute the population that Rimm
claimed to study.


That Rimm and Sirbu then submitted a grant to the DoJ that could be
used to bust the very people who were his subjects goes beyond any
breach of research ethics  that I can recall, ever, in the social
sciences.  This is not a minor lapse of ethics or an error in
judgment. It is a fundamental violation of the most basic principles
of the treatment of human subjects.


Sirbu acknowledges that he was not only fully aware of Rimm's
methodology, but that he would defend it. Hence, he was not unaware
of the population of BBSes from which Rimm drew his data.  From the
existing evidence, it is clear that Sirbu was the driving force
behind the DoJ grant that would put those subjects at severe risk. In
fact, the grant was *DESIGNED* to put those subjects at risk.


*QUESTION: How does Sirbu explain what appears to be a sanctionable
violation of ethics?


*QUESTION: In the (presumably required) Human Subjects application
for CMU, did Sirbu fully apprise the Human Subjects review committee
that, while there may be no "human subjects" in the proposed DoJ
grant study, the research was designed to put at risk subjects of a
previous (SURG) CMU funding of which he was the supervisor?


There is a curious footnote relevant to the grant application.  In
the GLJ article, Rimm devotes considerable space to describing
Amateur Action BBS, and calls the sysop, Robert Thomas, the Marquis
de Cyberspace (GLJ: 1912).  The propriety of the GLJ discussion has
been discussed elsewhere (eg, CuD 7.58). What has not been discussed
is Rimm's relationship with this research subject. Mike Godwin
contacted Thomas's wife, and reports part of the response:


      That Martin Rimm was a member of the Amateur Action BBS,
      that he quarrelled publicly and privately with Robert and
      Carleen Thomas about how they ran their BBS (among other
      things, he wanted them to change the way their BBS software
      kept track of downloads), that his messages to them after
      they refused to comply with his "suggestions" grew angry
      and threatening, that he declared publicly that he would
      not renew his membership at Amateur Action, and that he
      *did* renew his membership in February of this year.


Godwin also revealed that Thomas's wife produced the printout of a
message from Rimm to Thomas in which which Rimm tells Thomas how
much he admires him and how he hopes to be his "friend forever."
In July, Godwin asked Thomas's wife if there were any information
in her records pertaining to Rimm. Godwin summarized it on
The Well:


     For example, his application for a renewal
     of his AABBS membership lists the same street address as
     that of the "Carnegie Press." A different phone number,
     though -- this one doesn't get you that weird message from
     the phone company when you call it. Instead, it just rings.
     Credit card number with (as I recall) an April 96
     expiration date. Purchase on February 17 of a "six month"
     membership, which, according to Carleen, actually means
     he'll be current until August 17, 1995.


     Rimm first solicited membership in Amateur Action BBS in
     May of 1994.  This is interesting since he's told at least
     one person that he didn't even know Robert Thomas's name
     until July of '94 when Thomas was convicted on obscenity
     charges in Memphis. Since the application from Rimm used in
     May of 1994 has Robert Thomas's name and address listed at
     the top, this seems unlikely.


     Want to know the best thing about the '94 application form?
     It was *mailed* in. It's filled out in Rimm's handwriting.


If the records are accurate, Rimm, now involved in a funded study
designed to facilitate prosecution of active "pornographic" BBSes,
renewed his membership on the BBS that he described as the "market
leader in adult pornography" (GLJ, 1854).


Given Sirbu's professed close association with Rimm and the study's
methodology, it is inconceivable that he was not aware of how the BBS
data were collected.


*QUESTION: Why did Sirbu not intervene to assure that ethical
procedures were followed, given the evidence that they were not?


*QUESTION: Did Sirbu himself conceal information about unethical data
gathering?


There are so many questions that CMU's Committee of Investigation
could ask that only a portion can be suggested here.  Whatever the
answers to the above questions, it is clear that something rather
unacceptable occurred in the conduct of this research. The visibility
of the study and the use of the "findings" by policy-makers, which
was an explicit intent of the study, require a thorough airing of
these concerns.


                          AN AFTER THOUGHT


Here's why I continue to be concerned with the Rimm "Cyberporn" study
and Carnegie Mellon's handling of the investigation of it.


I teach research methods.  I teach methods to sociology students in a
senior capstone methodology course. I teach methods to graduate
students in a seminar that draws students from several disciplines.
In these courses, I include a strong ethical component.  I'm not a
dogmatic ethical purist, and I recognize the difficulty of walking
the thin line between "ought" and "ought not." But, there are two
fundamental principles I emphasize to students:  1) Always protect
research subjects from any harm that your research may cause, and 2)
Never deceive or lie to research subjects.  It appears that not all
at CMU share these precepts.


Tonight I began the ethical component of the graduate methods
seminar.  The course is comprised of Masters and Doctoral students
and Faculty. Each of the students has a topic, derived either as a
course project or from their thesis/dissertation work. One student
described a project that required "infiltration," deception of
informants, and role-playing to secure the confidence of subjects.


I thought of Rimm's study and the ethical message it would convey to
this student: Research that specifies deceit and leads to harm is not
only acceptable, but publishable  in a reputable journal.  "If they
can do it at CMU, why can't we do it at NIU?"


What can I say to the students and faculty about "real world" ethical
behavior? What can I say to the student who argues that it may be
acceptable to lie to subjects for the purpose of data gathering?  How
can I explain the proper role of a faculty research supervisor if a
faculty advisor at a major research institution violates fundamental
ethical precepts and the school seems to condone it?


If Carnegie Mellon University remains silent on the questions raised,
it will be complicit in a standard of research behavior that simply
cannot be condoned. How CMU responds to the individuals involved is
an internal matter that hopefully will be handled with compassion.
However, this does not preclude an explicit and unequivocal
statement, derived from a thorough investigation, that disavows both
the Rimm study and the research model on which it is based.