A Guide to Best Practices in Human Subjects Research

Institutional Review Board

Important Note

This document is intended to provide guidance for students and faculty who conduct classroom exercises that involve human subjects. Such exercises are presumed to involve only minimal risk to the subjects, that is, risk that is no greater than subjects would encounter in their everyday lives. If you anticipate that a planned classroom exercise will involve more significant risks, or if there are unique or idiosyncratic elements to your project that do not conform to the descriptions in this document, you should consult first with your departmental representative to the Institutional Review Board (IRB). If your department does not have a standing representative, you may consult with the IRB representative in a closely related department, or contact the Chair of the IRB at 577-3623 (rackoff@bucknell.edu).

Introduction

Faculty in a number of disciplines at Bucknell assign to their students classroom exercises that involve human subjects. The proper conduct of research with human subjects is governed both by federal law and University policy. As an extension of Bucknell's policies, the principles of best practice described in this document apply to all disciplines that use human subjects in classroom exercises, even those that may not normally come under federal review (e.g., artists interviewing passers-by, or historians conducting oral history interviews). It is the joint responsibility of the instructor and the student to be aware of and comply with the applicable portions of this document. Note that this guide outlines only the basic requirements for working with human subjects. Individual instructors are likely to have more detailed and specific guidelines that are appropriate to their particular disciplines.

What is the Institutional Review Board?

To ensure the protection of human subjects, federal law and institutional policy require that all human subjects research is reviewed and approved by an Institutional Review Board (IRB) before the research is initiated. The Bucknell IRB is composed of 9 members who meet on alternate weeks during the academic year. A meeting schedule is posted in Notes & Notices at the beginning of each semester.

More specific definitions are provided below for research, human subjects, and protection.

What constitutes research?

Under federal law, research is defined as any systematic investigation designed to develop or contribute to generalizable knowledge. This definition excludes classroom exercises that have student learning as their primary purpose. It would be inconsistent, however, for an educational community to have different guidelines that apply in varying contexts. Bucknell's IRB has thus extended the same principles for the protection of human subjects to classroom exercises. Since prior IRB review and approval is not required for classroom projects, it is the responsibility of the instructor to be sure that student projects meet the institution's ethical and scientific standards. Students and instructors may also consult with the departmental representative to the IRB, an individual who is trained and experienced in human subjects research issues within the discipline.

Some instructors voluntarily submit for IRB review and approval a "general protocol" describing a proposed classroom exercise with human subjects. This protocol establishes the boundary conditions for the class assignment. Although individual students and instructors may have some discretion in designing a particular project, they will be expected to operate within the limits established by the general protocol, which has been approved by the IRB. An example of a general protocol from the Department of Psychology is attached as an appendix to this document.

Resources related to human subjects research are available on the IRB website, and inquiries about classroom exercises may be directed to the Chair of the IRB, Dr. Jerome Rackoff, at 7-3623 or rackoff@bucknell.edu .

How are "human subjects" defined under IRB policies?

According to federal regulations, human subject means a living individual about whom an investigator (whether professional or student) conducting research obtains either

· Data through intervention or interaction with the individual; or

· Identifiable private information.

The term "intervention" refers to physical procedures (e.g., drawing blood in a medical experiment) as well as manipulations of the environment that may occur, for example, in a psychology or education research project. The term "interaction" refers to communication or interpersonal contact between investigator and subject. "Private information" includes information about behavior that occurs in a context in which an individual can reasonably expect that no observation or recording is taking place.

This seemingly straightforward definition can sometimes be confusing to apply in actual practice. For example, a number of disciplines employ naturalistic observation-simply going out into public places and observing the behavior of individuals in particular settings. This type of activity involves no intervention, and reasonable people cannot assume that they will NOT be observed when they are in public places. One could thus conclude that individuals engaged in public behavior are not really "subjects" under the federal definition. Indeed, federal guidelines include "observation of public behavior" in the category of activities considered to be "exempt" from further levels of IRB review beyond an initial assessment by an individual authorized by the IRB. At Bucknell, the departmental representatives to the IRB are authorized to grant exemptions, and students should consult with them regarding observational projects. If the course instructor is also the departmental representative to the IRB, students should consult with the designated alternate representative to the IRB.

The definition of "human subjects" is also difficult to apply when the focus of the study appears to be an organization rather than an individual, as often occurs in anthropology, sociology or management. Although individuals may be interviewed as part of the study, the data relate to the organization rather than the individual. Such studies would not be considered research with human subjects. Nonetheless, the disclosures of individuals who are interviewed in organizational studies can present other challenging ethical issues. Students should keep their instructors advised of any questionable issues that may arise.

From what are we protecting human subjects?

Human subjects must be protected from a variety of types of risks that can be associated with true research or classroom exercises:

· Physical Risks - Biomedical experiments are often accompanied by real risks to health and safety-adverse drug reactions, physical injury, complications of surgery, even death. (Physical risks are not common in classroom exercises, but they do occur from time to time.)

· Psychological Risks - Manipulating the variables in an experimental situation can expose subjects to a range of psychological reactions, including stress, embarrassment, boredom, etc. Personal histories and genetic pre-dispositions may cause some individuals to experience more pronounced psychological reactions than their fellow subjects.

· Legal Risks - The disclosure of information about illegal activities can place subjects at risk of prosecution, imprisonment, etc.

· Financial/Economic Risks - The inadvertent disclosure of information about personal health problems, for example, can put subjects at risk of losing health insurance coverage. Other disclosures could cause a subject's dismissal from his/her job.

The above list is not meant to be exhaustive. The important point is that experiments can pose very real dangers, and it is the responsibility of the investigator to design an experimental protocol that minimize risks.

How much risk is too much risk?

Institutional Review Boards must analyze the balance between foreseeable risks and anticipated benefits. The development of a new cancer drug promises very significant benefits to society, and measured risks may be justified. On the other hand, if a study is based on flawed science or unsound methodology, there can be no public benefit, and no amount of risk can be accepted. Classroom exercises fall somewhere in between these extremes. They do not contribute to generalizable knowledge, but they do have an educational benefit. The IRB recommends that students exercise prudence in designing educationally valuable classroom exercises that expose subjects to the minimal possible risk.

What is the meaning of minimal risk?

Human subjects activities are considered to be minimal risk if the risks to subjects are no greater than those they would encounter in everyday life. As we know, we are all subject to a variety of risks in our daily lives. Human subjects exercises should not increase that baseline level of risk.

What is informed consent?

Investigators must respect the autonomy of their subjects. Subjects have the right to make informed decisions regarding their participation or non-participation. For consent to be "informed," the experimenter must provide full and detailed disclosure of the circumstances of the study. The research community has developed a standard format for such "informed consent" (see below), which must be signed by the subject. The investigator must also offer the subject a copy of the signed consent form.

NOTE: Signed consent forms are not required for classroom demonstrations or exercises that use only students in the same class as subjects. The IRB recommends that instructors display prominently in their syllabi any obligations to serve as human subjects for classroom exercises.

Who can sign a consent form?

Subjects can legally sign their own consent forms ONLY if they are over 18 years of age. The participation of minors (e.g., for a project in a child development course) requires the signature of a parent or guardian, as well as the "assent" (agreement to participate) of the minor. Remember that many Bucknell students, particularly in the first-year class, are under 18. When using Bucknell students as subjects, it is generally impractical to get timely parental consent, and participation in classroom exercises should normally be limited to students over the age of 18 (see "components of informed consent," below).

Classroom projects should generally restrict their subject pool to normal, healthy individuals who are free to make their own decisions. Individuals with cognitive impairments, prisoners, and a number of other special populations receive additional protections under federal law. If access to these populations is essential to a particular classroom exercise, students should consult carefully with their instructor and the departmental IRB representative on the design of the project.

What are the components of informed consent?

· A statement that this is a Bucknell University classroom exercise.

· An explanation of the purposes of the exercise.

· The expected duration of a subject's participation.

· The anticipated number of subjects participating in the study.

· A detailed description of the procedures to be followed.

· A description of any foreseeable risks or discomforts to the subject

· A description of any benefits to the subject or to others (society, student learning) that may reasonably be expected from the research

· A statement regarding anonymity or confidentiality. If records identifying the subject will be maintained, indicate the extent to which these will be kept confidential.

· If deception is involved (see section on "deception," below, for more details), a statement to the effect that, "We cannot explain all of the details of the experiment to you at this time, but they will be explained fully at the conclusion of the experiment."

· An explanation of whom to contact for pertinent questions about the research (generally the student investigator and/or instructor), and whom to contact about research subjects' rights and research-related injury (Dr. Jerome Rackoff, Chair, IRB, c/o Office of VP Academic Affairs, 7-3623)

· A statement that participation is voluntary, refusal to participate will involve no penalty or loss of benefits to which the subject is otherwise entitled, and the subject may discontinue participation at any time without penalty or loss of benefits to which the subject is otherwise entitled.

· When Bucknell students will serve as subjects, the following statement should precede the signature block:

"In signing below, I affirm that I am 18 years of age or older."

This statement is unnecessary in situations where the subjects are clearly all of legal age (e.g., interviews in a nursing home).

· Dated signatures for subject and investigator.

What is the difference between active vs. passive consent?

The signing of an informed consent agreement constitutes "active consent." In some experimental situations, however, it is difficult or impossible to get signed consent forms, and "passive consent" may be appropriate (consult with the instructor and/or departmental IRB representative to be sure). With passive consent, one assumes that anyone who returns a survey form has passively consented to participating in the study project. Note that the experimenter is still responsible for disclosing all of the essential elements of a consent form in the initial communication. Subjects cannot make an "informed" decision to return a survey unless they have all the information that would normally be included in a signed consent form.

One form of passive consent that is NOT endorsed by the IRB is the "opt-out" consent. An example of this approach might involve sending a consent form home to parents with a cover letter that explains that their children will be participants UNLESS the parents return a signed form saying that they do NOT want their children to participate. Since one cannot be sure that the parents ever received the initial communication, this kind of consent cannot be justified.

What is "institutional approval"?

For class projects that use as their subjects individuals in organizations other than Bucknell (e.g., churches, public schools, or other universities), students should first seek permission from a responsible authority at the organization.

If the other organization is a university or hospital that has its own IRB, it is the obligation of the investigator to comply with all IRB policies of Bucknell AND the other organization. Depending upon how the other organization deals with classroom exercises with human subjects, this may mean submitting a proposal for review by the other IRB.

When the subject of study is an outside organization (rather than individuals within the organization), it is also normal practice to secure prior approval from a responsible authority. (See also discussion beginning on Page 2: "How are "human subjects" defined under IRB policies?").

If there are special circumstances that make it impractical or impossible to secure institutional approval, students should consult with the departmental representative to the IRB for guidance.

What is the difference between anonymity and confidentiality?

Students are often unclear about the differences between anonymity and confidentiality. If no names are collected on survey forms, the survey is anonymous, and respondents' identities are protected. If the experimenter can associate individual subjects with their responses, the study is NOT anonymous, and procedural steps must be taken to ensure that those responses remain confidential. Even the best procedures can fail, however, leading to the accidental disclosure of subject responses. The IRB thus recommends that exercises with human subjects be conducted anonymously whenever possible. Anonymous data collection is the ONLY acceptable method if the classroom project requires the collection of sensitive information. (Note that special precautions apply to e-mail surveys. See section below.)

Why are data collection procedures important?

Even anonymous surveys can be compromised by poorly designed data collection procedures. For example, if subjects turn in completed, anonymous survey forms directly to a survey assistant, that person can associate the name and face of the person with the survey responses. The survey is no longer anonymous. Ballot boxes and other procedures can be readily implemented to ensure that the researcher has no way to associate the subjects with their responses.

Why is deception such a significant concern of the IRB?

Deception involves withholding information from subjects that might affect their decision to participate in the study. The IRB regards very seriously any use of deception, since withholding information violates the fundamental ethical principle of autonomy. If we have respect for subjects as autonomous individuals, we also respect their right to a make a decision about their participation based on full information. Nonetheless, there are certain types of research that would be impossible without deception (e.g., fields such as social psychology), and deception is acceptable under federal regulations as long as appropriate protections are provided.

Deception occurs in varying degrees of severity. In its most benign form--incomplete disclosure--subjects are told the truth but not the whole truth. The only information that is typically withheld is the experimental hypothesis to ensure that subjects provide unbiased responses. Progressively more severe examples include (a) deceiving subjects about the purpose of the experiment, (b) deceiving them about the status of other individuals who they believe to be subjects (confederates), and (c) deceiving them about the status of individuals supposedly outside of the experiment (e.g., persons allegedly needing help in a study of helping behavior). The most extreme form of deception occurs when participants are not even aware that they are subjects until after the experiment has concluded. No classroom study involving deception at a level greater than incomplete disclosure should be conducted without prior consultation with the instructor and the departmental representative to the IRB.

The IRB recommends the following principles of best practice in studies involving deception:

· Deception should never be used in a classroom exercise if there is an alternate way of studying the research question without deception. One cannot expect students to imagine all possible ways that an exercise might be conducted, but the IRB expects that students, in consultation with their instructors, will make a good-faith effort to find alternatives to deception.

· Incomplete disclosure (to protect the research hypothesis) is acceptable as long as the project follows the practices outlined below.

· Deception that is more severe than incomplete disclosure should never be employed in a classroom project without prior consultation with the course instructor and the departmental representative to the IRB. (Note: If the instructor and the department's IRB representative are the same person, the instructor should consult with the designated alternate representative to the IRB.)

· Every experiment involving deception must include the following provisions:

o The consent form must advise subjects that they are not receiving all of the relevant information prior to the experiment, but they will be fully informed at its conclusion. The IRB recommends the following language: "We cannot explain all of the details of the experiment to you at this time, but they will be explained fully at the conclusion of the experiment."

o Subjects must receive a thorough debriefing at the conclusion of the experiment, including a disclosure of the deception and an explanation of why it was necessary for the experiment. A complete debriefing script should be approved in advance as part of the methodology of the study.

o To restore subjects' autonomy and control (that is, to restore the right to decide on participation based on full information), experimenters must, at the conclusion of the debriefing, offer subjects the opportunity to withhold the use of their data if they are unhappy with the deception.

What is different about e-mail surveys?

E-mail is a promising survey medium since it reaches quickly large numbers of potential subjects. It is a medium, however, that has methodological difficulties that can place subjects at considerable risk.

· All mail is traceable to one's e-mail address, so it is impossible to have anonymous e-mail surveys. (Technically, there are ways of sending unidentifiable messages through "anonymous re-mailers," but most subjects are unfamiliar with such resources. Further, even messages from re-mailers can be subpoenaed if a criminal action is suspected.)

· E-mail is an insecure medium, so the experimenter cannot assure confidentiality. Hackers and others can intercept and read e-mail messages, and messages reside almost indefinitely on the University's servers, from which they can be subpoenaed in criminal cases. In the aftermath of September 11, the United States has also signed an international treaty and enacted its own legislation (the US Patriot Act) to help combat terrorism. The new legislation gives authorities wider latitude to monitor electronic communications if links to terrorist activities are suspected.

How can e-mail cause problems for subjects? Imagine an e-mail survey that asked for sensitive information about family health history. If that information was inadvertently made public, it could prevent an individual with health problems from qualifying for life and health insurance. The IRB thus recommends the following principles of best practice for e-mail surveys:

· Preface your e-mail survey with a concise introduction that provides the basic information required in a signed consent form (see "components of informed consent," above). Remember that subjects responding to your survey will be considered to have given "passive consent" and that consent can only be "informed" if the subjects have complete information about your project. An example of such a consent paragraph follows:

I would appreciate your response to the following survey, which is being conducted as a Bucknell University classroom exercise for Psychology 0.7. We are interested in learning your attitudes towards____________. The survey should take only 5 minutes to complete. There are no foreseeable risks to your participation (other than the cautions about confidentiality noted below), and your responses will benefit the educational process in Psychology 0.7. For further information about this project, you may contact (name of instructor) at (campus phone number).

· Provide the following, brief, boilerplate paragraph after your introductory information:

" By replying to this survey, you are giving your informed consent to participate in this project. You also acknowledge that e-mail is not a secure medium, and that the confidentiality of your response cannot be guaranteed. If you wish to respond and are concerned about confidentiality, please print the completed survey and forward it anonymously via campus mail to __(name of contact person)__ at the following address: ___(campus mailbox)____."

· In general, avoid using e-mail as a medium for returning sensitive information. If your project absolutely requires sensitive information, use e-mail as a vehicle for quickly reaching a wide audience, but then ask all respondents to use the technique described above: print the completed form and return it to you anonymously via campus mail.

Some subjects will most certainly continue to respond via e-mail. This should not be problematical as long as all potential risks have been disclosed, since their passive (implied) consent will then be fully informed.

What about Web-based surveys?

Undergraduates conducting classroom exercises will generally not have access to the resources for Web-based surveys. Nonetheless, this is a medium that is widely used for a variety of types of human subjects research, and will no doubt become more accessible in the near future. Web-based surveys raise human subjects concerns that are similar to those for e-mail surveys. How can anonymity or confidentiality be assured? How can informed consent be given? In some cases, the nature of the technology provides novel solutions to these questions. For example, it is common practice to have readers click on a button to signify agreement with a Web policy. Federal and state governments have been quick to adopt the concept of online "signatures." These approaches could be adapted to provide consent for participation in a Web-based survey. Students contemplating Web-based surveys should discuss the project with their instructor and/or the departmental representative to the IRB.

J. Rackoff

Revised: 8/09/2005