Rationalizations by low-paid microworkers raise questions of ethics, rigor

Microworkers that researchers hire from online sites may feel conflicted by their low wage, leading them to both overestimate the importance of their work and put forth less effort, according to Penn State researchers.

In a study, researchers found that lower-paid Amazon Mechanical Turk microworkers—or Turkers—were more likely to reduce their efforts, as well as experience psychological states that could impact the results of the study. Microwork sites offer scientists and other researchers an inexpensive way to recruit participants, who often agree to be paid well below the minimum wage, added the researchers, who report their findings in an upcoming issue of Computers in Human Behavior, available online now.

“Researchers don’t want participants to have superficial involvement in their studies because, as a result, the rigor of these studies, and therefore the findings, could be compromised,” said S. Shyam Sundar, distinguished professor of communications and co-director of the Media Effects Research Laboratory.

Bingie Liu, a doctoral student in mass communications who worked with Sundar, said low pay may prompt a Turker to experience cognitive dissonance, which is a psychological state that arises when a person’s thoughts or beliefs are not aligned with their actions.

Liu said that one example of cognitive dissonance is a person who believes wasting water is bad for the environment, but then takes extra-long showers. In the case of Turkers, those performing tasks with too low a payment might experience cognitive dissonance because the compensation does not seem fair for the amount of effort expected of them.

“People might handle cognitive dissonance for this in a few different ways,” said Liu. “For example, if they experience cognitive dissonance about saving water, they might change their minds and say saving water isn’t that important, or they may try to align their actions with their beliefs and change their showering habits.”

People who have cognitive dissonance may take seemingly contrarian stands. For example, in this study, low-paid workers were more likely than high-paid workers to say their performance in the study was important, a sign that cognitive dissonance is at work. She added that this type of cognitive dissonance could cloud the participants’ perceptions during the study and, therefore, skew research results.

“Our data showed that overstating the importance of the study is a coping tactic because it is associated with more enjoyment and less tension in their minds while filling out the questionnaire,” said Sundar. “However, it is also linked to a tendency to please the researchers by providing answers that they think the researchers are looking for rather than honest, authentic responses.”

“A subject experiencing cognitive dissonance during a study is a problem for scientists because most research isn’t focused on cognitive dissonance—they’re studying other psychological processes, so researchers don’t want to bring in this irrelevant effect,” said Liu.

The researchers also said that low-paid microworkers tended to drop out of studies earlier without finishing the study and put forth less effort in their participation, such as responding with fewer words to open-ended questions.

Microwork sites are becoming increasingly popular as online laboratories for research, according to Sundar.

“There are hundreds of experiments going on at any given time on Amazon Mechanical Turk with thousands of participants from all over the globe,” said Sundar. “This is a big enterprise right now. It’s like a lab online, open 24/7, that is constantly operating.”

According to Sundar, while the study looked at how wages affect workers on microwork sites, the results may have implications for companies that outsource work to these sites.

“This study looks at the underside of what’s been described as an ‘online digital sweatshop’ and how those conditions might compromise what they do,” Sundar said. “On the one hand, they might not do the work you assign them to do properly, but on the other hand, there’s a real ethical issue because you’re putting people through all of these states of dissonance where they’re torn and they’re struggling with rationalizing to somehow make the process work for them.”

The researchers recruited Turkers who lived in the United States to take part in two experiments that took about 15 minutes to finish. In the first experiment, the researchers recruited 145 people, with one half of them paid $1.50 and the other half $0.50. In the second experiment, a total of 149 participants were recruited, with half being paid for $3.00 and the other half $0.25.

Source: Read Full Article