Project I: Informed Consent and Perceptions of Risk and Harm in Survey Participation

PRINCIPAL INVESTIGATOR
Eleanor Singer
Senior Research Scientist
Institute for Social Research
University of Michigan

CO-INVESTIGATORS
Frederick Conrad
Associate Research Scientist
Institute for Social Research
University of Michigan

Mick Couper
Senior Associate Research Scientist
Institute for Social Research
University of Michigan

Robert Groves
Director, Survey Research Center and Senior Research Scientist
Institute for Social Research
University of Michigan

Project Summary

Researchers have an obligation to inform participants in research of the risks of harm to which they may he exposed as well as of any potential benefits, and to do so in ways that facilitate comprehension of the information presented. This project uses laboratory research, web-based experiments, tests of informed consent statements with representative samples of the population, and a validation survey to answer the questions such as the following: What risk of disclosure is the public willing to accept? Is disclosure of some kinds of information considered more harmful than others? Are some intruders into the data perceived as more harmful than others? Do people perceive relationships between expected risk of harm on the one hand, and magnitude of harm and likelihood of disclosure on the other, in a manner consistent with the mathematical implications of those terms? How can researchers accurately inform participants about disclosure risks without unnecessarily deterring their participation? Our formulation of these research questions will be guided by the findings of the Project investigating statistical disclosure risks and limitations.

Specific Aims of the Project

Researchers have an ethical obligation to inform participants in research of the risks of harm to which they may he exposed as well as of any potential benefits, and to do so in ways that facilitate comprehension of the information presented. They must successfully communicate to participants the likelihood that information about them will remain confidential, or conversely, the risk that it will not, without unnecessarily deterring their participation. The need to sustain a high level of participation is crucial for survey research, because without it, the generalizability of the findings may he jeopardized.

The major aim of Project 2 in this broader program on the risks of human subject disclosure is to measure, evaluate and reduce the risks of statistical disclosure as a result of deliberate attempts to breach confidentiality. But reduction of objective risk is not enough. As Slovic, Fischhoff, and Lichtenstein (1982:464) point out. "assessment [of risk] is inherently subjective, and understanding judgmental limitations is crucial Researchers, that is, must learn how to communicate information about disclosure limitation in such a way as to reduce, rather than increase the fears of participants.

Although there is a relatively large literature on comprehension of informed consent statements primarily in medical research, past research on comprehension of confidentiality assurances is quite limited. The present study aims to augment this literature by answering the following questions:

1. What risk of disclosure is the public willing to accept?
2. Is disclosure of some kinds of information considered more harmful than others?
3. Are some intruders into the data perceived as more harmful than others?
4. Do people perceive relationships between expected risk of harm on the one hand, and magnitude of harm and likelihood of disclosure on the other, in a manner consistent with the mathematical implication of these terms?
5. How can researchers accurately in form participants about these risks without unnecessarily deterring them from participation?

Information derived front both Projects 1 and 2 is necessary in order to develop "best practices" of data dissemination, which is the goal of Projects 3 and 4.

The "Informed Consent and Perceptions of Risk" project will provide answers to the questions above by means of qualitative exploratory and quantitative experimental research. We will use Web-based experiments that permit graphical as well as verbal representations of risk, followed by tests of informed consent statements with representative samples of the general population and by a validation survey. In designing these experiments, we will draw on our previous work as well as oil other projects in the current program proposal that are investigating the actual risks of disclosure. Thus, for the first time, an investigation of informed consent procedures in social and behavioral research will be informed by statistical analyses of the actual risks of disclosure present in such research. In turn, our work will provide essential guidance for researchers struggling to balance accurate communication of the risk of harm with credible assurances of confidentiality.