bypassed. But this is rare, and it is clear that the researchers in the Cornell–Facebook study failed to obtain consent from the thousands of Facebook users who were subjected to the manipulation of their news feeds. Instead, the researchers took advantage of the fine print in Facebook’s data use policy to conduct an experiment without users’ informed consent. Even though the academic researchers collaborated with Facebook in designing the study, it appears that they obtained ethical approval only after the data collection was finished (Chambers, 2014).
Researchers have argued that informed consent is not required for research in online contexts in which the data can be considered to be in the public domain (Eysenbach & Till, 2001; Sudweeks & Rafaeli, 1996). And professional research associations occasionally deem informed consent unnecessary in cases where the scientific value of a research project can justify undisclosed observation. However, in cases where it cannot be legitimately argued that data are in the public domain or where data are in the public domain but are protected by copyright laws, participants’ informed consent to use such data must be sought.
Because the process of seeking informed consent is onerous and requires the creation and administration of an IRB-approved informed consent form, text mining researchers typically prefer to use data that are clearly in the public domain.
Manipulation
So far, we have assumed that the researcher is collecting unprompted user conversations (rather than prompted data, such as from interviews or questionnaires), but social scientists are beginning to collect users’ textual data after actively manipulating the online environment as a stimulus intended to assess reactions or responses. The Cornell–Facebook emotion study is an example of such research. Researchers could also prime users by, for example, introducing sexist, racist, or homophobic language into the online environment and then recording the responses of members of different communities. From an ethical standpoint, for this kind of experimental online research it is not sufficient for the researcher to anonymize participants’ names after the experiment has been conducted.
As always, in the case of manipulation, the best practice is for university researchers who plan to manipulate the social media environment in any way to consult with their IRB and for researchers in the industry to follow the regulations and guidelines of their respective professional associations.
Publishing Ethics
If you are thinking about graduate school or a career in research and teaching, you have many outlets available for publishing your research papers. You can publish your own work in specialized undergraduate research journals, present your work in undergraduate poster sessions at national and international academic conferences, and possibly upload your undergraduate honors thesis to your university’s digital research archives. You may also publish collaboratively, as research assistants or perhaps occasionally coauthors, with faculty members in research journals and conference proceedings. Whatever your specific goals, it is important to be aware of the many ethical pitfalls involved in scholarly publishing. In this section, we borrow liberally from research ethics scenarios (http://ethicist.aom.org/2013/02/ethics-in-research-scenarios-what-would-you-do) that were developed by management researchers Davis and Madsen in 2007. The following scenarios presented all represent ethics violations related to authorship and publishing, and they all represent patterns of behavior that occur quite often.
Scenario 1
In the first scenario, you have recently begun to work with a professor who is a productive scholar who has published in major journals for many years. But you have discovered that he has an unusual approach to research. He begins by gathering and analyzing data, which may include using a student’s data set, to see if the data have anything interesting to say. You have found that the professor often manipulates the data and changes the dependent variable to ensure a statistically significant result and increase the probability of a major publication.
Is this professor’s approach to research ethical? Why or why not?
Is there anything you could or should do as a student in this situation?
Scenario 2
In the second scenario, a professor has a long and impressive resume, but upon closer examination, you realize that many of her publications seem to be quite similar. One day you met with this professor and commented on her impressive body of work. She said that she never writes anything that doesn’t get as much ink and attention as possible. Among other things, she said that she may change the name of some of her papers to get them into conferences. She also claimed that she spends so much time gathering data that to be as productive as possible she must use the same data and theory in multiple published studies.
Can one plagiarize oneself?
How often can data be used ethically?
Can the same paper be submitted to a conference and for publication in a journal?
Scenario 3
In the third scenario, three graduate students were chatting one evening about their frustrations. Student A said that she had written a final paper for a course taught by her adviser. When she asked him if the paper was worthy of submitting to a conference, her adviser said that it is worth submitting and suggested the insertion of a few references. He then requested that he be listed as the paper’s coauthor. Student B then shared the story of his adviser’s request. He explained that as he wrote his dissertation he received wonderful support from his adviser, but once the dissertation was completed his adviser insisted on being first author on all publications coming out of the dissertation research. Student C trumped all three by saying that her adviser told her that he owns the data and all intellectual property (IP) coming from the dissertation because he consented to supervise her research.
How is authorship resulting from your research determined?
Who owns the IP and data from your research?
Scenarios 1, 2, and 3 all represent ethics violations related to authorship and publishing, and they all represent patterns of faculty behavior that occur quite more often than we might prefer to believe. What would you do in each scenario? In each of the scenarios, the faculty adviser has immense power over the student’s future academic career. An adviser who refuses to write a letter of recommendation for a student for a job or graduate school or who writes an unflattering letter can cause permanent damage to the student’s career. Davis and Madsen considered these and other ethical scenarios on their blog post, which is a very useful resource (http://ethicist.aom.org/2013/02/ethics-in-research-scenarios-what-would-you-do). You can also refer to the research ethics sites in the Web Resources and Further Reading sections toward the end of the chapter.
Conclusion
Choosing appropriate guidelines for text mining research can pose a serious challenge. Watson, Jones, and Burns (2007) suggested that it is impractical to suppose any single set of guidelines could address all issues concerning online research due to the diversity of online platforms and communities. Hair and Clark (2007) made the point that what might be considered ethical research conduct in one community might be considered unethical in others. In light of this uncertainty, it is critically important that you pay close attention to what is considered
