Survey Research and Ethics

Seymour Sudman, University of Illinois, Urbana-Champaign
[ to cite ]:
Seymour Sudman (1998) ,"Survey Research and Ethics", in NA - Advances in Consumer Research Volume 25, eds. Joseph W. Alba & J. Wesley Hutchinson, Provo, UT : Association for Consumer Research, Pages: 69-71.

Advances in Consumer Research Volume 25, 1998      Pages 69-71

SURVEY RESEARCH AND ETHICS

Seymour Sudman, University of Illinois, Urbana-Champaign

I am delighted to participate in this Presidential Session on Ethics in Consumer Research. Part of our mission as researchers is to socialize new researchers in the field as to what are the ethical issues that we face. I will attempt to highlight the issues facing survey researchers, but of course many of these same issues face other types of consumer researchers. As my framework, I shall be using the Code of Professional Ethics and Practices of the American Association for Public Opinion Research. Over the past several decades AAPOR has devoted significant amounts of time and thought to issues of how this code should be worded and enforced.

The code spells out a researcher’s professional responsibility to respondents, clients or sponsors, and the public. I’ll discuss each of these in turn.

RESEARCHER RESPONSIBILITIES TO PARTICIPANTS

Researchers have some general obligations to the people who provide data in marketing research studies. These include:

- Participants should not be harmed.

- Participants should not be deceived.

- Participation should be willing and informed.

- Data should be held in confidence.

Participants Should Not Be Harmed

The first obligation of researchers to participants is not to harm them in any way. Almost never does physical harm become an issue in survey research, although in a few health surveys blood samples may actually be taken. At an emotional level, the obligation not to harm participants means that they shouldn’t be embarrased, ridiculed, belittled, or generally subjected to mental distress.

Taken to the extreme, the obligation not to hurt participants might seem to make research impossible. The approach taken by most researchers who have pondered this issue is to apply two standards. First, researchers should minimize the potential harm of a research project (and maximize respondents’ convenience). This entails practices such as:

- Reminding research participants that they need not answer any question they prefer not to answer.

- Scheduling interviewing so as to minimize disruptions. Second, once the potential harm of a research project has been minimized, compare its risks to risks that people face in everyday life. If the risks posed by the research are no greater than those posed by everyday life, then the research is not viewed as posing serious ethical issues.

Ethical issues do arise when a person who is asked to participate in research does not instantly agree. Is it ethical to try to persuade someone to participate and, if so, how much effort is permissible before it becomes harassment?

The approach that many researchers have adopted is to distinguish between hard and soft refusals. A soft refusal is one that says "I’m too busy right now," or "This isn’t a good time." A hard refusal is one where the person says "No, I’m not interested." Hard refusals are not followed up, but soft refusals may be. Even for soft refusals, there will almost never be more than one follow-up effort.

For mail surveys, respondents who send back the questionnaire indicating that they don’t wish to participate are not followed-up, but those who don’t mail it back receive subsequent reminder mailings.

Participants Should Not Be Deceived

There are many ways in which researchers might deceive respondents. Is it ever ethically permissible to deceive participants in research? Several situations can be distinguished.

Sales prospecting and fund raising under the guise of surveys: Because so many people are willing to participate in research studies, some unscrupulous marketers and fund-raisers use phony research, especially surveys, for purposes unrelated to obtaining information. This is a clear violation of all ethical codes that deal with marketing research. If something is represented to participants as research, then it should be legitimate research, and the results should not be used for other purposes without the express consent of participants.

As an example of phony research, a major car company conducted a large telephone survey some time ago, indicating to respondents that the purpose of the survey was to measure their attitudes about cars and their buying intentions for market research purposes. This information was then turned over to local dealers who made sales calls to likely buyers. Not only is this an unethical deception of respondents, but such phony surveys ultimately lead to reduced cooperation on legitimate surveys, killing the goose that lays the golden eggs.

One of the most common uses of phony research these days is to raise money. Fund raisers have discovered that people are more likely to contribute money if they have a chance to express their views on issues related to the fund raising. Thus, a brief survey, usually of four or five questions, is included with a request for a contribution. The survey results are typically thrown away after the checks are cashed and no attention is given to them. Fund raising under the guise of doing research is deceptive and unethical.

Misidentification of the researcher or research sponsor: In my opinion, sponsor or researcher misidentification is acceptable if it serves legitimate purposes of interviewer seurity or data integrity and it does not encourage participation from people who otherwise would not participate. Consider a survey about car attitudes done on behalf of the Ford Motor Company. If respondents are told that Ford is the research sponsor, they may shade their answers to be more favorable toward Ford cars. This gives a legitimate reason not to disclose the research sponsor. At the same time, it seems unlikely that people who are willing to participate in a survey about automobiles would suddenly withdraw that participation if they knew that Ford was the sponsor. Therefore, I don’t see an ethical problem with failing to disclose the sponsor’s identity at the start of the survey.

If there is reason to believe that the sponsor’s identity will affect people’s willingness to participate in the research, then the sponsor should be identified.

Deceiving participants about the length of the task: Imagine a researcher who wants to do interviews that will take 90 minutes. The researcher doesn’t mention how long the interview will last during the introduction because of a concern that many respondents will refuse to cooperate if they know its length. In this case, this is deception by omission and is unethical.

In general, the length of an interview should be specified when seeking cooperation if it is possible to do so. For interviews where the length varies depending on the answers, it often is possible to specify a range.

Participation Should Be Willing and Informed

The whole ethical issue of how much information to give research participants so they can make informed decisions on whether to participate is a difficult one. Respondents should be given adequate information, but a researcher should not take 20 minutes to explain a ten minute questionnaire. Such detailed explanations increase the total time required from respondents with no clear benefits.

The general principle that guides informed consent is that the amount of information should bear some relationship to the risk involved in the study. Most survey research involves only small risks for the participants, and only minimal information need be given at the beginning. For example, a typical survey introduction would need only three kinds of information:

- Who is doing the study?

- What is the study about?

- How long will it take?

One might want to add :"Your participation in this study is, of course, completely voluntary. Feel free to refuse to answer any question that you feel is too personal." Such a phrase makes respondents’ rights more explicit, and has been found to reassure people and make them more likely to participate.

It is unethical to falsify the name of the research firm or the client. If the research is done in-house by a company, then that company should be willing to be identified. Some companies have established in-house research operations and given them different names to conceal the parent company’s identity, a borderline ethical behavior.

Data Should Be Held in Confidence

The final researcher obligation to participants is the obligation to hold data in confidence. Whether or not a promise of confidentiality is made, research participants have a right to confidentiality unless they specifically waive this right.

This means that data gathering organizations should routinely separate information that can identify a specific respondent from the information the respondent gives. This usually cannot be done at the time of a survey interview because it necessary to have identifying information for purposes of verifying that the interview was cnducted. It also may be necessary to call respondents back if interviews are incomplete or crucial answers are unclear. However, once verification and data editing are complete, identifying information should be separated from the answers and never supplied to the client or anyone else. In fact, the identifying information should be destroyed unless there are plans to conduct follow-up research in which the same respondents are reinterviewed. In that situation, the identifying information should be retained in a file separate from the data, and respondents should not be promised anonymity (though it is possible to promise confidentiality).

Research participants may waive confidentiality if they wish. In some research projects, participants are asked if they would be willing to be identified so that someone from the sponsoring company can talk with them to follow up their comments. Since such followups are done with the permission of participants, there is no violation of ethical principles.

A rare but troublesome problem in research is to have data subpoenaed for court cases. When data are subpoenaed, there are only limited legal safeguards for respondent confidentiality. If identifying information has already been routinely removed, of course, then respondent confidentiality can be maintained, but this information cannot be destroyed after being requested by the courts.

RESEARCH SUPPLIERS’ OBLIGATIONS TO CLIENTS AND SPONSORS

The survey research supplier is a professional who is ethically obligated to meet professional standards. These standards can be categorized into four main areas:

- Proper procedures should be used.

- Benefits of the research should not be over-promised.

- Clients’ information should be kept confidential.

- Results should not be distorted.

Proper Procedures Should Be Used

The AAPOR code includes the following statements:

We shall recommend and employ only those tools and methods of analysis which, in our professional judgment, are well suited to the research problem at hand.

We shall not select research tools and methods of analysis because of their capacity to yield misleading results."

It sometimes is difficult to separate poor competence from poor ethics, but researchers are ethically bound to do the best they can for clients, and to advise clients against inappropriate courses of action.

Problems most often arise when the client proposes procedures that the researcher believes to be inappropriate. Sometimes this is because the client really believes that the proposed method is better than the one recommended by the researcher. Other times, though, the client may be proposing a procedure that will yield a desired result.

Many research companies, faced with the possibility of losing the revenue from a project, will simply do whatever the client wants. However, an ethical researcher will attempt to persuade the client to change the design, and, if this is not possible, will refuse to do the study rather than be associated with an inappropriate project.

Do real-world researchers ever turn down business? Believe it or not, they do. Many ethical firms refuse to do work that they consider unprofessional. In the long-run, both they and their clients benefit from this behavior.

Benefits Should Not Be Over-Promised

In order to win projects, research companies sometimes promise results at a level of accuracy greater than is possible with currently available research techniques. This is unethical behavior. The AAPOR Code puts it this way:

"We shall be mindful of the limitations of our techniques and capabilities and shall accept only those research assignments which we can reasonably expect to accomplish within these limitations."

Clients’ Information Should Be Kept Confidential

Clients often need to reveal proprietary information about their marketing plans and product development to research suppliers in connection with projects. Such information could, of course, be valuable to competitors, and researchers have an ethical responsibility to keep it confidential.

It is unlikely that any supplier would be foolish enough to reveal proprietary information while still working with a client, since this would cause an instant dissolution of the relationship. Researchers are ethically bound, however, to keep former clients’ data confidential even if they are now working for a competitor. Again, this is simply good business. If a researcher reveals information about a previous client to a current one, the new client will be unwilling to trust the researcher with any proprietary information.

Results Should Not Be Distorted

Again the AAPOR code states:

"We shall not knowingly make interpretations of research results, nor shall we tacitly permit interpretations that are inconsistent with the data available.

We shall not knowingly imply that interpretations should be accorded greater confidence than the data actually warrant."

Note that this code not only requires researchers to avoid making false claims about their work, but also requires them to speak out when their client or someone else makes incorrect public statements about the data. This is not an easy to do, and not all firms do it, but ethical research firms do. The researcher has no ethical obligation regarding use of a study within the client’s own organization.

DISCLOSURE REQUIREMENTS WHEN RESULTS ARE PUBLICIZED

Earlier I noted that research data should be handled in such a way as to protect the confidentiality of clients and participants. This naturally limits the proper disclosure of methods and results. For example, a research company should not describe research that it has done for clients, or even disclose the existence of research, unless clients have given their permission.

However, if research results are publicized by the client an effort to persuade other people to take some action, then the client and the researcher have an ethical obligation to provide these people with enough information to evaluate the research. The confidentiality of individual participants should be sacrosanct, but the researcher or client should make available:

- A copy of the questionnaire

- A description of the sample design and execution.

- A description of where, when and how the data were collected.

- Data analyses relevant to the conclusions of the report.

I am not naive enough to believe that all survey researchers adhere to these ethical standards I have propounded. Some people will do anything for a buck or a publication, but there are many of the best and most successful who do worry about ethical issues, and I hope you will too.

----------------------------------------