Research Ethics in Victimization Studies

Abstract

Research ethics, especially the protection of research study participants, has been an enduring concern of investigators and scholars of victims of violence. Previous work has analyzed the overt and hidden risks of harm to study participants through poorly designed and implemented protocols, cross-cultural insensitivity, unjust exclusion of certain victim types, and neglect to ensure safety of research staff. This article extends this work by calling for greater attention to exploiting participants and victims as a class of persons. The need for this wider view derives from concern about the academic research environment with its emphasis on research, publication, and extramural funding. The authors argue that compliance with federal regulations and IRB directives is necessary but insufficient to conduct truly ethical research. Recognition of the unique pressures that researchers face in the context of the economically competitive university is required to delineate and manage exploitation risks. Such pressures have sometimes led to the rush from print to practices and policies with unintended harmful results for victims of violence as a class of persons. The authors suggest a number of strategies for researchers to define and manage the dangers of engaging in this type of unethical conduct.

Keywords

exploitation partner violence research principlism research ethics research with victims

Introduction

Curiously, victims of partner violence are not defined by Federal research regulations as a class of vulnerable individuals. The designation is limited to individuals who could likely face coercion, such as prisoners, pregnant women, fetuses, neonates, and children (U.S. Department of Health and Human Services, 2009). Yet many researchers examining violence victims recognize the need for special protections to reduce the risks associated with study participation for victims of violence (Buchanan, Fisher, & Gable, 2009). Consistent with this concern, the literature on research ethics among studies of partner violence and violence against women focuses on the direct participant protections needed during study implementation. Exemplary papers by Btoush and Campbell (2009) and Fontes (2004) present approaches to safely implement studies using basic protections for participants and researchers. Fontes (1998) has also incisively addressed the cross-cultural issues in conducting studies with victims of family violence.

Research with victims of violence is a Scylla and Charybdis of ethical problems that reveal tensions between research participant protection and victims’ opportunity to contribute to important research by participating (“Research Is Changing,” 2006). On one hand, avoiding this population for fear of causing harm deprives women of the opportunity to articulate their experiences in ways that might help reduce violence—clearly an ethical concern reflecting paternalistic overprotectionism (Beauchamp & Childress, 2009; Newman, Risch, & Kassam-Adams, 2006). In fact, it is clear that women, as a class of persons, have been unjustly excluded from biomedical and social science research in the United States (DeBruin, 1994; Sherwin 2005). On the other hand, research participation can result in harm to participants, sometimes outside of the direct awareness of the investigators.

IRBs routinely express alarm about social and behavioral studies that require violence victims’ participation, often with only limited or no empirical evidence to buttress their objections (Becker-Blease & Freyd, 2006; Newman, Walker, & Gelfand, 1999). Furthermore, IRBs sometimes forbid the conduct of meritorious scientific work (Singer & Levine, 2003). More importantly, the specific harms related to study participation have been empirically examined and findings generally support the idea that the risks of participation are modest (particularly in light of medical studies), with a few exceptions, including a small percent who report feeling distressed from the disclosures and a very small percent who regret participation (Fontes, 2004; Johnson & Benight, 2003). These effects can be moderated by careful attention to the participants’ emotional needs. For example, Hlavka, Kruttschnitt, and Carbone-Lopez (2007) found that debriefing female participants attenuated the unpleasant effects of recounting their victimization experiences. Given these findings, IRBs may be changing their hypercautious reviews (“Research Is Changing,” 2006). However, it remains unclear whether participant disclosure of violence victimization in research studies—particularly in the case of rape—is, in itself, therapeutic or stress inducing or both (Campbell, Ahrens, Sefl, Wasco, & Barnes, 2001).

In this article, we propose widening the frame of reference for approaching the ethics of research with victims of partner violence. Part of the motivation for this broader view of ethics is the changes that have happened in the academic research environment that mean more bold incentives for investigators to conduct research and secure extramural funding. We raise the question about the ethics of research with victims out of concern for whether the promotional and entrepreneurial characteristics of contemporary universities may influence research undertaken at any cost and without sufficient regard for scientific rigor. It also may stimulate research without due caution about work with vulnerable participants. The contemporary university is characterized as a business that competes for extramural dollars and even takes on other commercial interests such as hotels and catering facilities (Thomas, 2010). This commercialized academic environment becomes the context within which research is conducted, and it creates a situation that is ripe for exploitation of participants in order to enhance investigator tenure and promotion opportunities (Jordan, 2003; Price, 2010).

The core of research ethics is, of course, the protection of participants from harm. Much of the ethics literature regarding research on victimization among women presents guidance or ethical rules to follow in order to carry out participants’ protections. However, the specific focus on protections of participants leaves out a wide array of other ethical concerns that rules do not easily touch upon. The goal of this article is to provoke questioning and thinking about the many ethical issues that are embedded in victimization research in the context of contemporary academia and public policy environments.

We argue that the literature on ethics in victimization research often overlooks the broader ethical issues that affect the net outcome of findings that are disseminated to practice and policy. Although some authors have acknowledged this wider ethical framework, it has been largely unexplored as an ethical issue affecting research participants or the specific class of research populations (Tangney, 2000). This article examines research ethics in relation to the broader aspects of research activity that go beyond the implementational aspects, such as data management, analysis, interpretation of findings, and the offering of policy or practice recommendations. We argue that widening the lens of ethical analysis requires investigators to move beyond mere rule compliance in participants’ protections; researchers must view their entire effort through a perspective of ethical conduct. A brief review of ethical theory opens the door to understanding how the rule-based ethics of today is grounded in a theory that has limitations for the diversity of studies of victimization among women. The strengths of a virtues-based ethics theory may be most resilient in meeting the many ethical demands placed on contemporary investigators.

Widening the Lens

While the ethics literature classifies scientific misconduct as unethical (Institute of Medicine, 2002), its harms are typically cast as injuring “science.” We suggest that misconduct is unethical because it can exploit the class of persons who make up the target population for the study. The rationale for this broader perspective is that research findings incorrectly or superficially analyzed can result in harm to a class of persons quite independent of overt risk exposures to study participants. In essence, bad research can lead to mistaken beliefs, and mistaken beliefs can have untold consequences for victims in systems of care and protection. Thus, for the purposes of this article, misconduct may be seen as a form that exploitation can take.

The widened ethical perspective also includes implications of journal decisions about publishing victimization studies. There are potential ethics questions related to high-tier journals refusing to publish applied research that is “not sufficiently experimental” or controlled and, on the other hand, from narrowly focused journals publishing too freely work that is substandard. The other dimension that we address is the question about potential exploitation of participants when the research is done solely to advance a dissertation or to further an academic career without making a substantive contribution to the science on victimization.

We suggest that the frame of reference through which research ethics has been considered has been too often narrowly confined to the immediate problems of protections during study implementation and that investigators have frequently been reduced to rule followers in place of becoming ethical thinkers. In this article, we encourage investigators of partner violence victimization to a more imaginative application of ethical thinking to their research design, implementation, analysis, and dissemination.

Primary Ethics Constraints: The Current Regulatory Structures

The history of errors in conducting research with human subjects is vast and need not be fully repeated here. (For extensive historical discussion and analysis, see Childress, 2000, and Childress, Meslin, & Shapiro, 2005). Likewise, the documentation of usual and customary protections has been reasonably well spelled out in the literature (Amdur & Bankert, 2007; DuBois, 2008; Sales & Folkman, 2000).

Research protections for participants are regulated by the “Common Rule” articulated in 45 CFR Part 46 that also defines the functions of institutional review boards (IRBs). Institutional review boards carry out the oversight of research conducted with federal funds and in academic institutions (Penslar, 1995; U.S. Department of Health and Human Services, 2009). The traditional method for ethical review of studies involves the application of rules to specific cases; this approach is principle based (Childress, 2007). The ground for principle-based ethical analysis is found in the three Belmont principles: respect for persons, beneficence, and justice (Barnbaum & Byron, 2001). However, these three principles are not ranked hierarchically and research studies often create conflict and competition among the three principles when particular cases are under consideration (Veatch, 2005). Given the number of ethics problems that arise in research, it would appear that the regulatory process is effective only in detecting and addressing the more extreme or obvious ethical errors.

What Is Ethical thinking? The Role of Theory?

Much of the current regulatory structure for protecting human participants is grounded in a deontological theory of ethics. Most closely identified with Kantian principlism, this approach posits that the dignity of humans is found in following one’s inherent duty to behave morally. In Kant’s scheme, the moral status of any decision is assigned by testing whether it can and should be universally practiced by all persons. Those proposed maxims that pass this generalization test can earn the status of moral principles from which moral rules may be derived (e.g., “self-evident truths”). Deontological moral theory produces an ethics of strong principles where the moral actor is obliged to not deviate unless a countervailing moral principle justifying that rule breaking can be established (Childress, 1998).

Principles permitting and forbidding different types of research have resulted in codes and laws that address the protection of human research participants. Deontological principlism has contributed greatly to the development of contemporary bioethical principles such as “respect for all persons,” “do good and do no harm,” and “act justly” that have been the foundation for research bioethics (Pellegrino, 2005). Such ethical principles have shaped federal law designed to protect human participants, for example, the Common Rule (45 C.F.R. 46). This deontological approach was historically important to help steer science away from a purely utilitarian ethics, which justified a range of even sadistic methods as long as they produced outcomes that benefitted those supporting and doing the science (Lifton, 1986).

The unintended effect of the deontological approach was that it generated a system of constraints and rules that often rewards “rule following” as opposed to “thinking ethically” (Sieber, 2004a). It is ironic and perhaps tragic that Kant’s vision of the restless, ethical thinker who intellectually struggles with the dialectic of doing his or her duty has instead produced thoughtless, authoritarian persons who distribute permissions, rewards, and punishments on the basis of a set of preferred rules. This is especially problematic because the widespread prioritization of regulatory compliance over ethical quality is not empirically based. Only a handful of studies evaluate whether IRBs effectively engage in their mission of ethics deliberation (Candilis, Lidz, & Arnold, 2006; DeVries, DeBruin, & Goodgame, 2004; Sieber, 2004b; Taylor, 2007).

Furthermore, some observers critique deontological approaches that serve as the basis of contemporary bioethics as encouraging exclusion of the voices of others who have had no part in elucidating the rules; these voices belong to women, children, and other marginalized persons who belong to exploited or “culturally different” populations (Fontes, 1998; Yick, 2007). For such groups, moral experiences are embedded in the social structures and cultural contexts in which they struggle to survive (Farmer, 1999; Kleinman, 2006). These structures are often unrecognized as influencing ethical principles themselves, opening up a contemporary critique that questions the moral status of the process of how decisions are made about which axioms become principles, and which principles become law.

Principlism’s Practical Constraints for Research

For example, DuBois (2008) argues that regulations cannot replace ethical thinking because they are inherently constrained by a number of factors: (a) Regulations focus primarily on protection of research participants, sometimes at the price of neglecting equally important ethical questions, such as a “duty” to participate in research; (b) regulations are usually slow to respond to filling people’s needs in crisis situations (e.g., pace of HIV drug trials); (c) they provide minimum institution standards and do not explicate the range of standards that could be met; (d) IRBs and other oversight entities cannot possibly monitor all research and cannot replace the ethical decision making of the research team; (e) researchers who actively “partner” with participants can sometimes better determine the extent of “genuinely respectful, beneficial or harmful” (p. 20) and devise ethical approaches that more effectively serve the needs of participants and their communities than those prescribed by regulators; (f) the focus and reliance on rules may lead researchers to simply “jump through hoops” of regulatory mandates rather than strive for ethical outcomes; and (g) sometimes rules and laws lead to bad outcomes, and simply following these rules does not absolve researchers of the obligation for mature ethical thinking and foresight. DuBois is correct to affirm an earlier assessment by Kahn and Mastroianni (2001) that researchers and regulators should strive beyond a culture of compliance to reach a “culture of conscience and responsibility.”

Changing the Focus From Obeying Principles to Preventing Exploitation

Principlism’s attraction is that rule following allows researchers and institutions to believe they derive the benefits of rigorous, ethical thinking without needing to actually think about ethics. But we argue that the primary ethical activity for researchers should be to not only follow rules but also prevent the exploitation of their research participants or the class of persons they are researching, for example, victims of violence. The kind of exploitation suggested here is where research is being done primarily to obtain a publication or to advance the academic career of the investigator without generating meaningful new information about victimization. Admittedly, this process may be unintentionally exploitative, but it is so due to the indifference or inattention of investigators—yet another ethical problem.

Exploitation has been a primary concern in research with a wide range of vulnerable populations, including violence victims (i.e., groups beyond the regulatory definition of vulnerable populations; Beauchamp & Childress, 2009). Wertheimer identified 11 characteristics of exploitation. Exploitation does not need to be intentional but can even be inadvertent when self-interests are most prominent. The central characteristic of exploitation is that one party has clear benefit or gain and the other party may or may not benefit from the activity (Wertheimer, 2008). In addition, we often associate exploitation with intentional actions, but it can also be inadvertent. Although it is generally wrong to exploit others, the degree of wrongness is less clear. We raise this as a concern because the risk of distress from participation should mean that only truly beneficial research should be entertained. The vulnerability in this case arises from participants’ belief that their participation will lead to meaningful contributions to the cause of reducing violence against women. It can be an exploitation of the altruistic aims of participants.

Perspectives Necessary for Preventing Research Exploitation

It is almost natural for the ambitious investigator to conceptualize research participants as being “variable carriers” from whom the investigation will extract and analyze variance and covariance as units in a sample of N participants. Investigators are especially vulnerable to exploiting participants when they see people that way—a collection of independent, dependent, and control variables encased by skeleton, skull, and skin. Or as participant first, and then and only then, as persons in the world who merit some general interest. However, there can be an alternative starting point:

All human lives are played out against a particular cultural and historical background. The background is what gives meaning to the philosophical question, “How should I live?” and in particular to a subset . . . [that] consists of questions like “What gives my life meaning?” . . . The answers to these questions, such as they are, I will call “final justifications.” . . . All the notion of final justifications requires is that we think of a life in its entirety, with its parts connected in some sort of unity. Just as a thread may not be held together by any single fiber, a life might not be held together by any single narrative. As Wittgenstein [1953, sec. 67] puts it . . . “the strength of the thread does not reside in the fact that some one fiber runs through its whole length, but in the overlapping of many fibres.” (Elliott, 1999, p. 137)

Wittgenstein’s (1953) insight appears to be a point of origin, or even a method, that would prevent exploitive research design. If the researcher cared to imagine the daily life of, for example, a young woman who is being stalked, a penniless, battered mother of three small children living in a shelter, or a woman lying in a doorway after being raped and dumped in the street, these images might evoke the kind of care and curiosity that would lead to more fruitful scientific concerns embedded in a fundamental desire to not make their lives any worse. When done in an informed and precise way, especially in the design of nonintervention research, it could assist investigators to approach the bioethical ideal of “equipoise,” which is the highest standard for exposing others to risk.

This narrative approach is quite consistent with another ethics theory that would greatly benefit researchers and participants, virtue ethics. In light of the unranked principles presented by Belmont, Beauchamp and colleagues (Beauchamp, Walters, Kahn, & Mastroianni, 2008) have argued for an ethical theory that relies on particular characteristics of the informed moral agent, as opposed to rule followers:

Virtue ethics . . . has practical value in that a morally good person with right desires or motives is likely to understand what should be done, to perform required acts, and to form moral ideals than is a morally bad or indifferent person. A trusted person has an ingrained motivation and desire to do what is right and to care about whether it is done. Whenever the feelings, concerns and attitudes of others are the morally relevant matters, rules and principles are not as likely as human warmth and sensitivity to lead a person to notice what should be done. From this perspective, virtue ethics is at least as fundamental in the moral life as principles of basic obligation. (pp. 18-19)

This seems to begin an adequate description of the required qualities for researchers seeking to avoid exploiting victims of violence.

One fundamental criticism of virtue ethics is that even persons of good moral character sometimes make bad mistakes or even act wrongly (Oakley, 1998). However, it seems unreasonable to argue that a person of good character who recognizes the ubiquity of human error would simply assume that he or she held all the answers. Rather, it is far more likely that persons of good character would carefully deliberate rules and guidelines that represent “lessons learned” by previous moral actors instead of solely relying on their own intellects. This integration of virtue ethics with the judicious use of laws, principles, regulations, and working knowledge of the outcomes of other “hard cases” would be the ideal approach for ethical persons. And this would seem even more likely to be the description of the “ethical researcher,” who is schooled from the earliest days of his or her training to grasp the fact that human beings are quite vulnerable to egregious errors in perception, information processing, reasoning, and decision making, and therefore, require rules, tools, guidelines, “peer review,” and other mechanisms to compensate for the deficits of the individual intellect (Clark, 2008; Einhorn, 1986; Kleinmuntz, 1990).

There is, however, an additional and significant caveat to the reliance on virtue-based investigators to carry out ethical research. Appiah cites studies conducted in the 1970s showing that even virtuous people act less morally and more selfishly when the immediate situation is either less rewarding or more stressed. Seminarians, being told they were late for an appointment, were far less likely to offer assistance to a stranger in need, and people who found a dime in the change return of a public phone were more likely to offer assistance to a stranger than those who did not find a coin (Appiah, 2008). It is the unmentionable “context sensitivity” of virtue that raises the concern with academic research. The change in the culture of universities to become entrepreneurial revenue centers has had an impact on the integrity of research that is still being studied (Bok, 2003). The entrepreneurial university creates environments in which promotion and tenure are affected by extramural funding and the number of publications. Virtue ethics, therefore, must also recognize its constraints.

Embedding Ethics in Research Design

Mere compliance with IRB or Common Rule regulations may miss the boat of genuine ethical practice of research. The rule-based approaches that are widely endorsed for participant protections may evade the more significant ethical issues that arise from research on this topic. We argue that “virtuous” investigators interested in preventing exploitation of participants, and victims of violence in general, explore embedding ethics in every step of study development, implementation, analysis, and dissemination.

For example, Logan, Walker, Shannon, and Cole (2008) demonstrated a way of using ethics to actually guide the design and implementation of a large-scale study of women with protective orders. Although the research team had an articulated approach that involved the wide use of both key informants and victims with protective orders, there were many unanswered questions about how recruitment and even interviewing could be done in such a way as to incorporate victim and key informant ideas about the processes and interview contents. Thus, they started with pilots to test out thinking about recruitment procedures, interview approaches, and even securing reactions of potential participants to the core instruments for data collection. The intent of this process was twofold: (1) Study personnel became sensitive to the perceived risks and possible benefits as seen by community participants, and (2) the efficiency and specificity of data collection were improved by drawing on viewpoints of potential participants about the meaning of variables and the ways in which more valid data might be obtained. While this approach focused on preventing exploitation, it also highlighted Belmont principles of benevolence and autonomy, both of which drew the research team into close accord with the safety and confidentiality interests of the potential participants. The approach also netted an improved research design and implementation with a 97% follow-up completion rate.

Buchanan, Gable, and Fisher (2009) and Fisher and Goodman (2009) recommend a “goodness-of-fit ethics” for research with high-risk populations. Researchers should attend carefully to vulnerable populations as defined by federal research regulations. But the duty cannot end there since there are other populations who are vulnerable and merit protection, such as victimized women. Instead of developing a simplistic and misleading classification of those who are vulnerable and those who are not, these authors suggest that researchers define vulnerability in a new way that emphasizes the interaction of the population’s characteristics with the research context and design. For purposes of clarity we have taken the authors’ recommendation and configured it as the formula, Pc x Rd = V, where participant characteristics interacting with research context and design estimates the probability of population vulnerability. Specifically, researchers can derive Pc through an extensive review of epidemiological, descriptive, and longitudinal research that examines the population under consideration. Pc can then be carefully considered, for example, when designing the informed consent procedures for a study. Although this sometimes means “going beyond” the duties prescribed by the IRB, it may often be, nonetheless, an ethical requirement (e.g., Iltis, 2006).

We suggest a slight modification to the formula: Pc x Pe x Rd x Re = V.

Although the notation is consistent with Buchanan, Gable, & Fisher’s (2009) conceptualization, we recommend introducing Participant environment (Pe), and Researcher environment (Re), as these both emphasize a bit more strongly the impacts of local, environmental factors that affect participants, as well as turning attention to the researcher’s academic and scientific contexts (Re) that can drive research design and contribute to participant vulnerability. How might this be effectively operationalized?

Logan, Walker, Jordan, and Leukefeld’s literature review of more than 2,000 sources in the area of women and victimization culminated in a conceptual model that describes four categories of significant social psychological factors contributing to victimization, mental health, and substance abuse: specifically, lifestyle factors, social factors, internal contextual factors, and victimization manifestations/vulnerabilities (Logan et al., 2006, Figure 8.1, p. 197). Turning to the literature specific to the study population and developing a conceptual model that addresses those factors could create a fine-grained description of the extant knowledge of Pc. Contextual factors of importance (Pe) can also be discovered in many published papers, especially those that convey detailed understanding of the participant’s neighborhood and local culture (Dunlap, Johnson, & Randolph, 2009). Estimating the impacts of research design (Rd) can be accomplished through careful review of human participant protection literatures, especially where risk of harm for the population under consideration is adequately explored (e.g., Buchanan, Gable, & Fisher, 2009). Finally, researcher contexts (Re) can be explored through a truthful examination by the research team about the pressures and incentives experienced by individual researchers as well as by the research team. By factoring the interactions delineated in this equation, researchers are likely to detect even more subtle forms of exploitation risk.

Research Questions and Exploitation

A naive or poorly formulated research question may pose an ethical problem if (a) it narrows down to inappropriate results, (b) it so closely replicates existing research that it would make the study redundant, or (c) it is unanswerable and, thus, would expose participants to a meaningless activity and needless risk. We suggest that these errors, which may be typically conceived of as products of naivety or lack of well-grounded education, may actually be unethical. They might be at a minimum exploitive because they would subject participants to some degree of risk with no gain in public knowledge and perhaps gain only for the investigator. A fourth error is to frame a research question in a “loaded” manner that will inevitably unfairly disparage one class of persons (often labeled by cultural, racial, or gender characteristics) in comparison to another group (Fontes, 1998). “IQ studies” have historically demonstrated this error (Gould, 1981).

Literature Review

A comprehensive, formative, and analytic literature review is a daunting task. Logan, Walker, Jordan, and Leukefeld (2006) conducted an integrative review of more than 2,000 articles and texts, and Btoush and Campbell (2009) found more than 2,400 research articles on the violent victimization of women. It is, therefore, inevitable that an empirical researcher will review literature selectively. To some extent, the selection will be guided by specific research interests, but even within this more narrowed focus, the literature will likely be complex, extensive, and conflicted about central findings. Thus, the investigator must decide what to include and what to exclude in the review. Novice researchers, eager to bolster their advancement into tenured or at least tenure-track positions, may be easily tempted to slant the reviews to support the need for their particular studies. The boundary between this as a legitimate scientific method and a self-serving shortcut is difficult to discern.

Research Methods, Analysis, and Exploitation

Research method is one of the treasured competencies of investigators. This is where there are complex interactions between the demands of the science and all the ethical issues around participant protection.

There are important ethical concerns even around the presentation of findings and the choices to be made about analytical approaches. For example, a researcher might discover bivariate associations that support a hypothesis where a more complex analysis controlling for other variables would not support that finding. Many papers, particularly epidemiological studies of partner violence victimization, rely on these levels of analysis, and the scientific community may accept this due to limitations in the data set or other factors. However, when possible, the more thorough analytical approach will be a more ethical solution because it comes closer to truth. Although the ethics literature examines the intentional distortion of findings, it is less clear how limited analyses should be interpreted from an ethical perspective.

Meltzoff (1998) notes that such errors create great potential to distort findings, and although they might not constitute outright data falsification or fabrication, they do “skirt” the border of unethical research conduct by misleading readers. He suggests willful contribution to error as seen in these examples: (a) claiming to hold “other things constant,” but failing to do so in the administration of instruments or the conduct of interviews, or even arranging conditions to favor the research hypothesis; (2) tampering with random assignment to favor the research hypothesis; (3) dropping participants after consent for no valid reason and then not reporting on the criteria for doing so or providing a description of the drop-out subsample; (4) creating and grouping categories to favor post hoc data analysis that favors the research hypothesis but is not consistent with the theoretical ground of the study; (5) tampering with rating or judging procedures by influencing the decision processes in favor of the study hypothesis; (6) developing criterion measures that are biased against a particular cultural, ethnic, or racial group and then using the data to demonstrate profound intergroup differences; (7) altering the number or sequencing of trials to favor the study hypothesis; (8) using qualitative data to distort or reverse unfavorable statistical findings; and (9) using graphs, tables, or pictures to bias the reader toward accepting the study hypothesis in the face of nonsignificant statistical findings.

Clearly, there are many temptations to cut corners or to nudge method in order to reach desirable, publishable results. These may never harm any specific research participant but can be considered exploitative of the class of persons for whom the study claims to be advancing scientific understanding

Examples of Ethical Problems by Study Types

Of the various victimization studies to get published, epidemiological studies of prevalence are among the most common and it is easy to see why; they are less costly and less complex to execute than intervention studies. The successful pursuit of an evidence-based practice can require years of study, multiple attempts at finding the best fit of intervention to population, and complex longitudinal designs to measure effects. On the other hand, cross-sectional epidemiological studies of prevalence are, by comparison, easier and faster to complete, particularly when the population is actually a subpopulation such as a class of clinic patients or members of a health service organization.

At one level, concern about this biasing in favor of less complex research is ill-placed due to resource limitations and the embarrassingly small number of promising interventions to test. However, at another level, there are at least two potential ethical problems: (1) the risk of mischaracterizing victim populations that can result from the casual use of subpopulation studies—thus, a justice concern, and (2) the risk of exploiting easily accessible populations for studies that are essentially replicative and add little or nothing to the general understanding of the prevalence of victimization among women. As to the first concern, careful attention to the analysis can blunt the possibilities of mischaracterization of victimization and related characteristics. The type of study that is of greatest concern here might be one where a health data set is examined for victimization rates and crude associations with health problems. Given the wealth of information about the relationship between victimization and health problems, it is unlikely that yet one more study of this will add anything meaningful to the field or to the treatment received by victimized women. In a worse-case scenario, the superficial analysis without controlling for other factors, may result in misleading associations about health and violence victimization.

Little has been written about the ethics of misleading findings. But, for example, do we routinely find studies controlling carefully for predisposing characteristics for a variety of health problems that have wide ranging other health effects, such as smoking tobacco and obesity and their contribution to health findings apart from victimization? Do we routinely find researchers carefully examining subsets of poverty conditions among victims that have been shown to be associated with profound health disparities independent of other characteristics?

It is one thing to criticize such research for its failure of scientific merit; it is quite another to suggest that it is unethical solely because its results might be clouded by factors not considered in the analysis. However, our earlier concern places a different ethical frame on such practices. Was the study carried out this way simply because it was easy to do with an available data set from a (presumably) consented sample? Should researchers consider whether the possible findings actually add to the body of useful knowledge as an ethical obligation, not merely as a pragmatic consideration of publishability?

Clinical Studies

Despite their relative disparagement in the social sciences, it is a great advantage to a field for researchers to develop a literature of rigorous clinical studies. Such studies allow the field to benefit from qualitative and experience-close approaches that can enrich theoretical and empirical work (Platt, 2007). Again, exploitation of individual participants can occur when such accounts violate the participant’s anonymity. If cases are written up from a clinical practice, exploitation can occur when a clinician researcher advances his or her agenda by offering therapeutic assistance only in exchange for clients’ consent to participate—a form of coercion. Finally, such research can ultimately backfire and harm a class of persons even if the case studies were used originally for advocacy purposes. Stark (2007) critiques the clinically based research that served as the basis for “battered women’s syndrome” (BWS). He argues that the quick translation of what could best be described as useful but preliminary clinical studies into a legal theory adopted by many court systems ultimately harmed many battered women charged in criminal cases. He contends that the portrait of the battered woman developed by these studies became the evidentiary standard even though ongoing empirical research revealed that battered women had a far greater variance of presentation. Thus, many battered women were unable to meet the criteria for BWS because they did not look like the women described by clinical researchers who advanced the profile.

Evaluation of Victim Services and Treatment Outcome Research

Evaluative research may use client-level data to measure program effort and outcomes (Royse, Thyer, Padgett, & Logan, 2006; Waxman, 1996). In evaluating programs that serve victims of violence, there may be covert or subtle coercion for women to participate. The coercion may come not from the investigator but from the program that is eager to demonstrate its value to a funding source. Thus, the risk for exploitation may be very high in these situations. In evaluation studies where there is no evidence of refusals to participate, there may be cause to question whether exploitation or coercion occurred. Also, there may be complex questions about confidentiality, particularly in studies that use follow-up designs where client-level contact data are necessary. In addition, there are always concerns about what to offer clients in the way of debriefing or referrals to other services when doing violence victimization research where sensitive disclosures are made. Again, from our perspective, these risks can be managed, but the ethical concern grows in relation to the limitations of the research itself. In other words, all things considered equal, a student research project with victims that has little or no contribution to broader knowledge has greater potential for unethical exploitation than evaluative research carried out to genuinely inform a public funding source.

Electronic Devices

The plethora of electronic data collection devices now on the market has greatly increased the ability of investigators to conduct applied research in community settings. However, there are few sources of privacy and confidentiality provisions regarding the use of research data on such devices. For example, consider the possibilities for personal, identified data to be unethically disclosed when identifiers and clinical data are in the same table on a portable device used in the field or identifiers are in separate tables but linkable to clinical data due to linkage tables being on the same portable device. The number and type of portable devices are ever increasing. At the time of this article, there are PDAs, netbooks, laptops, I-Touch, I-Pad, and cell phone devices that can be used to collect personal research data. Whereas harm resulting from careless use of portable data collection devices may be seen as a technical problem, again, we stress that a careful embedding of ethics in the total study would be a likely way to prevent or correct this problem. Portable devices can be used in electronic data collection studies to ensure confidentiality using even participant consent signatures, which are then encrypted to ensure voluntarism (Cole, Logan, & Walker, 2011).

Web-Based Research

The convenience of web-based research makes it attractive for a wide range of investigators. However, there are subtle and covert risks that should be considered. For example, many web-based survey instruments maintain the participant responses on servers not under the control of the investigator. For anonymous surveys, this is seen as a minor risk to confidentiality breach. However, investigators need to assure themselves that the survey program does not retain the Internet protocol (IP) addresses or email addresses of participants. Voluntarism of participation is presumed because participants can disconnect at any time. However, it is what is going on behind the scenes in the survey that can cause the risk to confidentiality. As stated earlier, these risks are manageable and justified as long as the research is of genuine scientific merit.

The Role of Scientific Journals

There is a significant gap in the literature on research ethics in regard to the role of the reviewers and publishers of journal articles. Anecdotally, researchers may comment on the weaknesses of certain journals or a certain article; however, it is another matter to question the ethics of those who are instrumental to dissemination. There are several factors relating to the ethics of article review and publication. The scientific standards of publications are used to select the types of articles that will be acceptable by journals. Many of the “high-impact” journals place a high premium on experimental design studies. Applied partner violence studies, as a class, virtually never appear within their pages. Hence, one sees numerous articles that examine issues that are sometimes peripheral to violence, for example, those using college students in laboratory conditions with exposure to stress exercises or subjected to brief trials testing performance of exercises. Victim advocates, clinicians in the practice environment, and violence researchers are unlikely to ever benefit from such studies due to the artificial settings and conditions that many experimental designs require.

The major medical journals seem biased in favor of clinical trials and research articles that test the clinical effectiveness of treatments and other clinical interventions. Violence victimization studies rarely feature these topics or research methods. In response to barriers to publication in these “mainstream” medical and psychological journals, narrower specialty journals have been developed that focus on partner violence and its related applied research. However, this literature tends to circulate among violence victimization specialists, so the opportunity for wider dissemination is limited. This raises the ethical question of whether the scientific bar has been set in such a way that vital information about victimization is blocked from key provider communities. Furthermore, the segregation of research knowledge on victimization into more focused journals has ethical implications for the larger practice communities who may have limited knowledge or access to this information. This barrier may result in health and mental health disparities among women seeking care for a variety of problems that arise from their victimization. The barrier to information being transmitted to health and mental health providers may constitute an ethical problem beyond mere editorial preferences.

Staff Training

Both authors are members of their university’s medical and nonmedical (behavioral and social science) IRBs and participate regularly in research ethics consultations. Protocols commonly give little evidence of being informed by ethical thinking even as they comply with the basic rules for protections using “template” language supplied by the IRB. We question whether “training” is what is needed to remedy this problem; perhaps education is the greater need. Training can inform rule application and form completion; education drives the ability to think and problem solve. Every study is unique (or should be), and, thus, the ethical problems that can arise will be unique. The solution to unique conditions is not rule following, but adaptive thinking. Competent education about the needs and risks attendant to research populations is critical to meeting the complex ethics issues that can arise (Ellsberg & Heise, 2002; Monahan, Appelbaum, Mulvey, Robbins, & Lidz,1993; Sullivan & Cain, 2004). The climate of overworked principal investigators working with limited research budgets under challenging deadlines can encourage merely obligatory regulation compliance at the expense of humane and ethical imagination and concern for others. These researcher environments make this a particularly vulnerable area for exploitation of violence victims as well as for those low on the academic hierarchy. One way to address this concern is for IRBs to develop ethics consultation services to help investigators think through the many study-specific design and ethics problems that can arise (Beskow, Grady, Iltis, Sadler, & Wilfond, 2009).

Conclusion

We have argued that ethical research with women who are victims of violence requires more from researchers than only complying with IRB rules or proscriptions. Indeed, the federal regulations that guide IRBs allow for greater degrees of flexibility than researchers typically recognize, and this flexibility suggests that researchers need to be more imaginative about their ethical responsibilities (Sieber, 2004b). Instead, we argue for a virtue ethics approach that cultivates the development of ethical researchers who take a more sophisticated, moral look at the entire enterprise of a research study, from conceptualization of the question to presentation of findings in refereed journals. This does not mean ignoring IRB rules of participant protection, nor does it mean throwing out the baby with the principlist bathwater. Virtuous researchers will use a pluralistic spectrum of approaches, with the ultimate rationale of preventing exploitation of their participants. Narrative and goodness-of-fit ethics approaches appear especially promising for this type of research. More importantly, we have also recommended a widened ethics lens to view the research participant and the class of persons who might also be exploited—in this article, women who are violence victims.

We also suggest that different steps in the research process and different kinds of research studies may introduce particularly problematic dimensions that can lead to exploitation. The development of empirical findings that are then translated into practices and policies—sometimes very quickly—places a particularly grave responsibility on the shoulders of researchers in this field. Although such applications can enhance the success and stature of the researcher, history teaches that such glory is often short-lived. Life is brief, but art (and science) are long, and a mature field acknowledges the need to wait for the replication o0f findings within similar studies and the convergent validity that comes from confirmations from other types of studies. For example, the advocates’ rush for the courts to adopt BWS has backfired, as subsequent research demonstrated that BWS represented only a subpopulation of victims (Stark, 2007); many other battered women who have deserved consideration for mitigating or even exculpatory circumstances have been denied. Thus, an entire class of persons have been harmed because of researchers’ eagerness to claim their findings as the last word.

We have also pointed out the “third rail” of research ethics that may be the most important determinant of whether researchers exploit others intentionally or unintentionally, that is, the temptations to which even virtuous researchers may fall prey as a result of their career pressures and other situational constraints.

In closing, we realize that we have not exhausted all dimensions of exploitation risk, and we recognize that many others have been raising the same concerns over time. However, we hope that this article once again draws attention to a discussion that must be sustained to be effective. Fontes (1998) stated the challenge very well when she observed over 10 years ago that

Even with the best of guidelines and human subjects review boards and consultants, researchers ultimately face ethical issues alone with their consciences. Many ethical decisions will be based on the amount of overnight tossing-and-turning that a researcher can tolerate. Therefore, all researchers are morally obligated to think deeply about ethical issues, to discuss these issues with others, and to keep their moral compasses delicately tuned, even if all researchers’ compasses do not point precisely in the same direction. (p. 53)

Footnotes

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

The authors received no financial support for the research, authorship, and/or publication of this article.

References

Amdur

R. J.

Bankert

E. A.

(2007). Institutional review board: Member handbook. Sudbury, MA: Jones & Bartlett Publishers.

Appiah

K. A.

(2008). Experiments in ethics. Cambridge, MA: Harvard University Press.

Barnbaum

D. R.

Byron

(2001). Research ethics: Text and readings. Upper Saddle River, NJ: Prentice Hall.

Beauchamp

T. L.

Childress

J. F.

(2009). Principles of biomedical ethics (6th ed.). New York: Oxford University Press.

Beauchamp

T. L.

Walters

Kahn

J. P.

Mastroianni

A. C.

(2008). Contemporary issues in bioethics (7th ed.). Belmont, CA: Thompson Wadsworth.

Becker-Blease

K. A.

Freyd

K. A.

(2006). Research participants telling the truth about their lives: The ethics of asking and not asking about abuse. American Psychologist, 61, 218-226.

Beskow

L. M.

Grady

Iltis

A. S.

Sadler

J. Z.

Wilfond

B. S.

(2009). Points to consider: The research ethics consultation service and the IRB. Ethics and Human Research, 30, 1-9.

Bok

(2003). Universities in the marketplace: The commercialization of higher education. Princeton, NJ: Princeton University Press.

Btoush

Campbell

J. C.

(2009). Ethical conduct in intimate partner violence research: Challenges and strategies. Nursing Outlook, 57, 210-216.

10.

Buchanan

Fisher

C. B.

Gable

(Eds.). (2009). Research with high-risk populations: Balancing science, ethics, and law. Washington, DC: American Psychological Association.

11.

Buchanan

Gable

Fisher

C. B.

(2009). Best practices for responding to threats of violence in research ethically, and legally. In Buchanan

Fisher

C. B.

Gable

(Eds.), Research with high-risk populations: Balancing science, ethics, and law (pp. 233-252). Washington, DC: American Psychological Association.

12.

Campbell

Ahrens

C. E.

Sefl

Wasco

S. M.

Barnes

H. E.

(2001). Social reactions to rape victims: Healing and hurtful effects on psychological and physical health. Violence and Victims, 16, 287-302.

13.

Candilis

P. J.

Lidz

C. W.

Arnold

R. M.

(2006). The need to understand IRB deliberations. IRB Ethics & Human Research, 28, 1-5.

14.

Childress

J. F.

(1997). Practical reasoning in bioethics. Bloomington, IN: Indiana University Press.

15.

Childress

J. F.

(1998). A principle-based approach. In Kuhse

Singer

(Eds.), A companion to bioethics (pp. 61-71). Oxford, UK: Blackwell.

16.

Childress

J. F.

(2000). Nuremberg’s legacy: Some ethical reflections. Perspectives in Biology & Medicine, 43, 347-361.

17.

Childress

J. F.

Meslin

E. M.

Shapiro

H. T.

(Eds.). (2005) Belmont revisited: Ethical principles for research with human subjects. Washington, DC: Georgetown University Press.

18.

Clark

J. J.

(2008). Complex thinking for wicked problems: An application of Sharon Berlin’s analysis of dichotomous thinking [Special issue: Honoring the scholarship of Sharon B. Berlin]. Social Work Now, 39, 38-48.

19.

Cole

Logan

T. K.

Walker

(2011). Social exclusion, personal control, self-regulation, and stress among substance abuse treatment clients. Drug and Alcohol Dependence, 113(1), 13-20.

20.

De Vries

DeBruin

D. A.

Goodgame

(2004). Ethics review of social, behavioral, and economic research: Where should we go from here? Ethics & Behavior, 14, 351-368.

21.

DeBruin

(1994). Justice and the inclusion of women in clinical studies: An argument for further reform. Kennedy Institute of Ethics Journal, 4, 117-146.

22.

DuBois

J. M.

(2008). Ethics in mental health research: Principles, guidance, and cases. New York: Oxford University Press.

23.

Dunlap

Johnson

B. D.

Randolph

(2009). Ethical and legal dilemmas in ethnographic field research: Three case studies of distressed inner-city families. In Buchanan

Fisher

C. B.

Gable

(Eds.), Research with high-risk populations: Balancing science, ethics, and law (pp. 201-230). Washington, DC: American Psychological Association.

24.

Einhorn

H. J.

(1986). Accepting error to make less error. Journal of Personality, 50, 387-395.

25.

Elliott

(1999). A philosophical disease: Bioethics, culture, and identity. New York: Routledge.

26.

Ellsberg

Heise

(2002). Bearing witness: Ethics in domestic violence research. Lancet, 359, 1599-1604.

27.

Farmer

(1999). Infections & inequalities: The modern plagues. Berkeley: University of California Press.

28.

Fisher

C. B.

Goodman

S. J.

(2009). Goodness-of-fit ethics for nonintervention research involving dangerous and illegal behaviors. In Buchanan

Fisher

C. B.

Gable

(Eds.), Research with high-risk populations: Balancing science, ethics, and law (pp. 25-46). Washington, DC: American Psychological Association.

29.

Fontes

L. A.

(1998). Ethics in family violence research: Cross-cultural issues. Family Relations, 47, 53-61.

30.

Fontes

L. A.

(2004). Ethics in violence against women research: The sensitive, the dangerous, and the overlooked. Ethics and Behavior, 14, 141-174.

31.

Gable

(2009). Legal challenges raised by nonintervention research conducted under high-risk circumstances. In Buchanan

Fisher

C. B.

Gable

(Eds.), Research with high-risk populations: Balancing science, ethics, and law (pp. 47-74). Washington, DC: American Psychological Association.

32.

Gould

S. J.

(1981). The mismeasure of man. New York: Norton.

33.

Hlavka

H. R.

Kruttschnitt

Carbone-Lopez

K. C.

(2007). Revictimizing the victims? Interviewing women about interpersonal violence. Journal of Interpersonal Violence, 22, 894-920.

34.

Institute of Medicine. (2002). Integrity in scientific research: Creating an environment that promotes responsible conduct. Washington, DC: National Academies Press.

35.

Johnson

L. E.

Benight

C. C.

(2003). Effects of trauma-focused research on recent domestic violence survivors. Journal of Traumatic Stress, 16, 567-571.

36.

Jordan

K. A.

(2003). Financial conflicts of interest in human subjects research: Proposals for a more effective regulatory scheme. Washington and Lee Law Review, 60, 15-109.

37.

Kahn

J. P.

Mastroianni

A. C.

(2001). Moving from compliance to conscience: Why we can and should. Archives of International Medicine, 161, 925-928.

38.

Kleinman

(2006). What really matters: Living a moral life amidst uncertainty and danger. New York: Oxford University Press.

39.

Kleinmuntz

(1990). Why we still use our heads instead of formulas: Toward an integrative approach. Psychological Bulletin, 107, 296-310.

40.

Lifton

R. J.

(1986). The Nazi doctors: Medical killing and the psychology of genocide. New York: Basic Books.

41.

Logan

T. K.

Walker

Jordan

C. E.

Leukefeld

C. G.

(2006). Women & victimization: Contributing factors, interventions, and implications. Washington, DC: American Psychological Association.

42.

Logan

Walker

Shannon

Cole

(2008). Combining ethical considerations with recruitment and follow-up strategies for partner violence victimization research. Violence Against Women, 14, 1226-1251.

43.

Meltzoff

(1998). Critical thinking about research: Psychology and related fields. Washington, DC: American Psychological Association.

44.

Monahan

Appelbaum

P. S.

Mulvey

E. P.

Robbins

P. C.

Lidz

C. W.

(1993). Ethical and legal duties in conducting research on violence: Lessons from the MacArthur Risk Assessment Study. Violence and Victims, 8, 387-396.

45.

Newman

Risch

Kassam-Adams

(2006). Ethical issues in trauma-related research: A review. Journal of Empirical Research on Human Research Ethics, 1, 29-46.

46.

Newman

Walker

E. A.

Gelfand

(1999). Assessing the ethical costs and benefits of trauma-focused research. General Hospital Psychiatry, 21, 187-196.

47.

Oakley

(1998). A virtue ethics approach. In Kuhse

Singer

(Eds.), A companion to bioethics (pp. 86-97). Oxford, UK: Blackwell.

48.

Pellegrino

E. D.

(2005). Some things ought never be done: Moral absolutes in clinical ethics. Theoretical Ethics & Bioethics, 26, 469-486.

49.

Penslar

R. L.

(1995). Research ethics: Cases and materials. Bloomington: Indiana University Press.

50.

Platt

(2007) Case study. In Outhwaite

Turner

S. P.

(Eds.), The Sage handbook of social science methodology (pp. 100-118). Thousand Oaks, CA: SAGE.

51.

Price

(2010). Sins against science. Monitor on Psychology, 41, 44-47.

52.

Research is changing IRB views about PTSD studies. (2006, May). IRB Advisor, pp. 51-53.

53.

Royse

Thyer

B. A.

Padgett

D. K.

Logan

Y. K.

(2006). Program evaluation: An introduction (4th ed.). Belmont, CA: Thompson Higher Education.

54.

Sales

B. D.

Folkman

(2000). Ethics in research with human participants. Washington, DC: American Psychological Association.

55.

Sherwin

(2005). Belmont revisited through a feminist lens. In Childress

J. F.

Meslin

E. M.

Shapiro

H. T.

(Eds.), Belmont revisited: Ethical principles for research with human subjects (pp. 148-164). Washington, DC: Georgetown University Press.

56.

Sieber

J. E.

(2004a). Empirical research on research ethics. Ethics & Behavior, 14, 397-412.

57.

Sieber

J. E.

(2004b). Using our best judgment in conducting human research. Ethics & Behavior, 14, 297-304.

58.

Singer

Levine

F. J.

(2003). Protection of human subjects of research: Recent developments and future prospects for the social sciences. Public Opinion Quarterly, 67, 148-164.

59.

Stark

(2007). Coercive control: How men entrap women in personal life. NY: Oxford University Press.

60.

Sullivan

C. S.

Cain

(2004). Ethical and safety considerations when obtaining information from or about battered women for research purposes. Journal of Interpersonal Violence, 19, 603-618.

61.

Tangney

(2000). Training. In Sales

B. D.

Folkman

(Eds.), Ethics in research with human participants (pp. 97-105). Washington, DC: American Psychological Association.

62.

Taylor

(2007). Moving beyond compliance: Measuring ethical quality to enhance the oversight of human subjects research. IRB: Ethics & Human Research, 29, 9-14.

63.

Thomas

(May, 2010). What are universities for? Times Literary Supplement, 7, 13-15.

64.

U.S. Department of Health and Human Services. (2009). Code of Federal Regulations: 45 Public Welfare, Part 46, Protection of human subjects. Retrieved from http://www.hhs.gov/ohrp/humansubjects/guidance/45cfr46.htm#46.103

65.

Veatch

R. M.

(2005). Ranking, balancing or simultaneity: Resolving conflicts among the Belmont principles. In Childress

J. F.

Meslin

E. M.

Shapiro

H. T.

(Eds.), Belmont revisited: Ethical principles for research with human subjects (pp. 184-204). Washington, DC: Georgetown University Press.

66.

Waxman

H. M.

(1996). Using outcomes assessment for quality improvement. In Sederer

L. I.

Dickey

(Eds.), Outcomes assessment in clinical practice (pp. 25-33). Philadelphia: Lippencott Williams & Wilkins.

67.

Wertheimer

(2008). Exploitation in clinical research. In Emmanuel

E. J.

Grady

Crouch

R. A.

Lie

R. K.

Miller

F. G.

Wendler

(Eds.), The Oxford textbook of clinical research ethics (pp. 201-210). New York: Oxford University Press.

68.

Wittgenstein

(1953). Philosophical investigations. New York: Macmillan.

69.

Yick

A. G.

(2007). Role of culture and context: Ethical issues in research with Asian Americans and immigrants in intimate violence. Journal of Family Violence, 22, 277-285.