mTurk and HIT Mores

Mechanical Turk is an interesting artifact that's become one of the tools in the top drawer of researchers.  Initially started as Amazon's way of crowdsourcing tasks that required creativity, intelligence, and autonomy, it's become something much more than that.  

While mTurk has successfully globalized a labor pool, it's also reduced compensation rates for time at task.  Julian Dobson's recent article is worth a look, and describes something much more insidious than merely access to a larger pool of subjects.  While Dobson quotes estimates from Panos Ipeirotus that between $10 and $150 million in business runs through the mTurk site, and into skilled workers, Dobson also suggests that this compensation is far below the level guaranteed by labor law. 

Some tasks, Dobson notes, quite generously offer a whole $0.50 for a 150 word essay, while, in my own cursory review found such an outstanding deal quite rare.  The average from what I saw was more along the lines of a 500 word essay for $0.02.  

At some point there has to be a discussion about exploitation.  Is it fair to say that exploitation isn't bad if it's done for a higher purpose?  Or is that argument, itself, flawed?  Is exploitation limited to actions forced on an individual?  At what point can we say a behavior is elicited rather than a choice?  

The ethical dilemma is often sidestepped in research.  The potential gains seem to outweigh the risks, though the cost of exploiting a labor pool is quantifiable.  Festinger's research on dissonance suggests only one possible cluster of outcomes.  This notion is often forgotten, or summarily dismissed as not worth entertaining when considered against the potential gains.  

Scientists need to face their own image in the mirror on this issue.  Ethical guidelines are there for a reason, and little consideration is being paid to the possible affect research using exploitative tactics is having on the population, much less the example it sets for commercial interests.  

The largest question, and the most poignant that I can muster, is why am I the one making this argument?  As David Resnik of the NIH, says in his article, "a person who makes a decision in an ethical dilemma should be able to justify his or her decision to himself or herself, as well as colleagues, administrators, and other people who might be affected by the decision."

Again from Resnik, "When conducting research on human subjects, minimize harms and risks and maximize benefits; respect human dignity, privacy, and autonomy; take special precautions with vulnerable populations; and strive to distribute the benefits and burdens of research fairly." This point should be ruminated on in the larger academic community.  With only a minimal view of research and how it treats its participants, even I'm uncomfortable with the recent trend towards mTurk as a source of research subjects, and not just because of the tragically low rate of compensation.  Debriefing as well becomes problematic when subjects are not on hand.  

Exploitative practices in finding, and funding research subjects are inherently dubious.  This area of research needs to be at least discussed, if not debated in a larger forum.  Compensation, just as deceit, needs to be discussed within the construct of its inherent ethical considerations.  mTurk and like solutions should only be utilized as a method of last resort.