Situation
J.D. Brighton has studied a small community for some time using surveys and focus groups and recently published a paper that showed there were significant differences in perceptions of police behaviors in different parts of the community. The local police chief did not believe the research outcomes, so he requested the data to do his own analysis. (Actually, he hired, Emma Gustafson, a sociologist from the local college to re-analyze the data.) Dr. Brighton believed that if she gave the data to the chief (or Dr. Gustafson) that she would lose access to the parts of the community that felt the police were not fair. It would effectively prevent her from continuing research in the community.
Brighton was also concerned that the police chief or the analyst could recognize some respondents by carefully analyzing the demographic characteristics of the respondents. The police chief wanted to replicate the study but to fully replicate the study, it was critical that all demographic data be available to the chief. In addition, Brighton’s Institutional Review Board approved the research on the condition that anonymity be protected. Although there was no information in the data set that directly identified anyone, she believed that an intense analysis of the demographic characteristics might make it possible to identify at least some of the research participants.
Questions
- Should Dr. Brighton give the data to the police chief if she knows that it will prevent her from further research some of the communities she studies? How can the researcher be sure that it will prevent further access to the communities?
- Should the researcher give the data to the police chief if the researcher suspects the chief might be able to identify some respondents? Why would the researcher think the chief and/or his paid analyst would not preserve the confidentiality of the data?
- Which principle should Dr. Brighton violate – data sharing or the possibility of violating promises of confidentiality?
Discussion
An important norm in science is that data and research procedures should be shared so that other researchers can replicate the research to determine if they achieve the same outcomes. It is generally agreed that a researcher should have sole access to the data s/he gathered until a major paper is released from the study. At that time, other researchers should be able to use the data to see if the results can be replicated.
There are two reasons why researchers would not want to share their data. First, the data might not be fully analyzed or there might be need to continue the research process after the researcher does the initial analysis. Second, the confidentiality of the research participants might not be protected by subsequent researchers. Both are good reasons for not sharing data. At the same time, researchers recognize that bad procedures and personal biases can influence the processes and analyses of research data. Therefore, if data are not shared, then there are no checks on the outcomes.
Researchers must balance their need to protect those who provide information to them along with the norm of data sharing in science. Suppose, in the case described above, the chief was willing to take positive actions to remedy the problems if he could have a “second opinion.” Should the researcher prevent the good that might come to the communities by not sharing the data and procedures?
