The ethics of metaphor as a research tool

Abstract

The interpretive and subjective nature of qualitative research has led to growing utilization of arts-based strategies for data collection, analysis and dissemination. The defining characteristic of all such strategies is that they are largely subjective and intended to invoke personal responses in the ‘audience.’ Following that direction, many qualitative researchers are using metaphor to capture themes emerging from their analysis. In this article, we explore ethical aspects of using metaphor in describing results of qualitative health research and illustrate some of the complexities using a case study of research conducted by one of the authors. Our analysis is designed to sensitize researchers and ethics reviewers to some unique ethical issues inherent to this approach towards data analysis and presentation. Issues related to participant dignity, respect and vulnerability led us to suggest that researchers should take these points into consideration in designing their research and seeking informed consent. Metaphors can be linguistic devices, but also conceptual aids that help develop patterns in analysis or that facilitate re-interpretation. However, there is a thin line between artistic licence for better expression and distorting the participants’ actual experience and meanings. Researchers, and reviewers, must be aware of the danger to participant dignity and integrity when aesthetics overshadow actuality. The use of metaphor may also trigger tensions between researchers and participants, especially if member checking is used. The implications of participant withdrawal must be considered and conveyed to ethics reviewers and participants. It is important to have a plan in place for dealing with some of these issues. These should be detailed in the proposal and communicated to participants. Institutional research ethics boards should, on their part, be prepared to ask questions if such details are lacking in the proposal.

Introduction

The interpretive and subjective nature of qualitative research has led to the growing utilization of arts-based strategies for data collection, analysis and dissemination (Cox et al., 2010). Momentum in Canada, and abroad, is gaining for the field of arts and health, both clinically and research-wise (Cox et al., 2010).

Arts-based strategies in research come in myriad forms: theatre (Conrad, 2006; Cox et al., 2009; Rossiter et al., 2008;); poetry (Furman, 2006); photography, drawing and sculpting (Deacon, 2000; Hermen, 2005); photovoice, dance and song (Cox et al., 2010); as well as creative writing and narratives (Nisker and Daar, 2006), all of which require some way of presenting concepts in a form that is intended to resonate strongly with the viewer. Metaphor is often used in such a way and is generally understood to be an acceptable approach to presenting findings in text-based research (Aita et al., 2003; Anderson, 2008; Carpenter, 2008; Deacon, 2000). However, one could argue that metaphor, defined as a ‘… figure of speech, replacing one idea or object with another to suggest an analogous relationship’ (Carpenter, 2008: 274)’, is effectively present whenever one concept is replaced with another in the interests of achieving vividness or clarity. Metaphor is employed in diverse approaches to arts-based research, the defining characteristics of which are subjectivity and an intent to invoke personal responses in the ‘audience’. As such, metaphor in arts-based research of all types is intended to achieve the overarching aims of qualitative research, that is ‘… to understand how people think about the world and how they act and behave in it’ (Canadian Institutes of Health Research et al., 2010: 135).

The move from ‘pure science’ to ‘arts-based’ methodologies raises some interesting ethical issues for researchers, as well as research ethics boards. Most of the dialogue around research ethics has approached the question from the more traditional ‘pure science’ perspective, where objectivity, extraction from context, and distance of researcher from subject are hallmark values. Less has been written about ethical issues arising from engaged, subjective, up-close-and-personal qualitative research, and the deliberate attempt to provoke a subjective response in the consumer or audience.

In this article, we explore ethical aspects of using metaphor in describing the results of qualitative health research and illustrate some of the complexities using a case study of research conducted by one of us (KPM). We selected metaphor because it has been manifestly equated to language use generally (Lakoff and Johnson, 1980) and the nonverbal arts (e.g. auditory or visual arts) specifically (Hausman, 1989). This equation implies, at the least, that all arts-based strategies contain metaphorical aspects (Nuessel, 1990). The ethics of metaphor as a research tool will directly inform and illuminate the ethics of most arts-based research strategies.

Our analysis is designed to sensitize researchers and ethics reviewers to some unique ethical issues inherent to this approach towards data analysis and presentation. First, we present our approach to this ethical analysis by confirming the general values of research ethics. Second, we describe the use and implications of metaphor in research. Finally, we provide specific directions for ethics reviewers faced with this form of arts-based research.

Approach

Ethics is about the values of human conduct (Sherwin, 1992). International, as well as national, codifications of values, commonly expressed as ethical principles, direct the ethical conduct of health research. After World War II, the Nuremberg Code was developed by the military courts trying Nazi physicians. It contained ten principles for ethical human experimentation, which can be grouped under three major values: respect for autonomy, avoiding unnecessary harm, and ensuring that benefits outweigh risks (Office of Human Subjects Research, 2011). The World Medical Association introduced the Declaration of Helsinki in 1964, which was most recently confirmed in 2008, as the guiding principles for medical research involving human subjects (World Medical Association, 2008). Critical values connect to respect for autonomy and privacy, the import of independent research ethics review, the promotion of health, and the need for minimal impact on participants. The 1978 US National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research report, the Belmont Report, described the three major principles of research ethics to be respect for persons, beneficence and justice (National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research, 1978).

These initial codes were influential in the historical development of national codifications or regulations around the conduct of publicly funded research. In Canada, the values directing research ethics governance are captured in the Tri-Council Policy Statement: Ethical Conduct for Research Involving Humans (TCPS2) (Canadian Institutes of Health Research et al., 2010). Three core ethical principles form the basis of the TCPS2: respect for persons, concern for welfare, and justice. In the United States, Title 45, Part 46 of the Code of Federal Regulations (the so-called Common Rule) describes the protection of human subjects in federally funded research. Values related to respect for persons, privacy, non-maleficence, vulnerability and justice are particularly evident.

From this analysis, it is noted that respect for persons (which subsumes autonomy), promoting benefits, minimizing harms, and justice are the four common societal values directing the ethical conduct of human research. Because of the uniqueness of health research, we delineated three further, underlying values: dignity, integrity and vulnerability. Respect entails dignity, but dignity further clarifies why there is the need for respect. Through beneficence and non-maleficence, the integrity of participants is promoted, but, in recognizing integrity, the goal of balancing harms and benefits is unequivocal. Finally, we isolated vulnerability from justice. Concern for justice relates to equity and fairness in all aspects of research, including the inherent vulnerability of participants. Vulnerability in research is universal and unique, as well as relevant at group and individual levels, and may be particularly important in the context of research directed or framed by the arts.

Ethics of metaphor in research

‘The essence of metaphor is understanding and experiencing one kind of thing in terms of another’ (Lakoff and Johnson, 1980: 5). Via substitution, metaphor permits displacement and expansion (Carpenter, 2008). Metaphors can be linguistic devices, but also conceptual aids that help develop patterns in analysis or that facilitate re-interpretation (Anderson, 2008; Carpenter, 2008; Deacon, 2000).

In demonstrating that metaphor is ubiquitous in everyday language for common experiences, Lakoff and Johnson (1980) argued metaphors are pervasive and inescapable. We agree with the inevitability of metaphor, but we argue that, in qualitative research particularly, there are two levels to metaphor use. The first level, as described by Lakoff and Johnson (1980), is baseline, unconscious and emphasizes common experiences (amongst and between researchers, participants and academia). The second level is explicit, purposive, and focuses on uncommon experiences. The second level would include researchers who thoughtfully use metaphor as a research tool to describe, analyse or disseminate the uncommon experience at the heart of their research.

Our questions about using metaphor emerged during data analysis in a study designed to examine human, legal and ethical perspectives on the hospital-to-home transition process for young, ventilator-dependent children (Manhas, 2011). Research methods included interviews with family, homecare professionals, hospital professionals and government representatives. The empirical element of the study was conducted using grounded theory methodology (Glaser, 2004; Glaser, 2007; Glaser and Strauss, 1967). The study also involved legal analysis of policies and case law on the division of parental and public responsibility and ethical analysis of the qualitative, legal and literature findings.

The analytic use of metaphor arose organically during the concurrent collection and analysis of the interview data and gradually became directive to the analysis. The overarching metaphor that seemed best to capture the hospital-to-home transition process was a rowboat journey along a river. Smaller themes that emerged fit well with this conceptualization. For example, the sharing of responsibility through the transition could be understood as rowers in a boat, each pulling with different amounts of strength and direction at different times, with the helmsman holding ultimate responsibility for the path taken. In hospital, or the first leg of the river journey, the helmsman was the physician; at home, or the second leg of the river journey after the watershed of physically moving the child home, the helmsman was the mother.

The metaphor provided structure to the data, which facilitated analysis by promoting understanding and illuminating issues faced by families in a way that would not have been recognized without this artistic tool. This use of metaphor was purposive and explicit for understanding the uncommon experience of transitioning from hospital to home. However, use of metaphor generated some particular ethical concerns. Carpenter suggests that metaphor risks ‘oversimplifying phenomena and depicting the complex as trite’ (Carpenter, 2008: 280). We confronted the question of whether the metaphor was illuminating or instead acting to mask the actual nature, goals and challenges of the hospital-to-home transition process. Was the metaphor more ‘fanciful’ than actually reflective of the process? There is a thin line between artistic licence for better expression and distorting the participants’ actual experience and meanings. Disrespect could accompany distortion. Research ethics obliges the researcher to consider whether aesthetics overshadows actuality. Participants or ethics reviewers could argue that metaphor, or other arts-based strategies, affront participant dignity by altering experience.

A related problem is that with such an arts-based strategy comes subjectivity, which is generally appreciated in qualitative circles but not as appreciated in the broader research community. As a powerful representative tool, metaphor ‘can both reflect and reinforce dominant socio-political views and values’ or militate against them (Carpenter, 2008: 281). If the values being expressed through metaphor are not congruent with the values of participants, there is a danger that participants’ integrity may be undermined. The researcher must be alert to the possibility that his or her personal values may be dominant in the analysis, and may be marginalizing or trivializing actual views of participants. The argument that researchers aim for subjective and situated understandings cannot divest researchers of the obligation not to harm participants and to consider any potential for harm. The study under discussion was grounded in feminist ethics methodology, so attempts were made to identify assumptions and potential biases. However, it is possible that unrecognized biases or ideologies swayed the framing and content of the metaphor. Again, this concern connects to participants’ dignity in that the researcher must ensure that participant experience is not obscured.

A contrary argument is that as long as participants are aware that the research will involve subjective interpretation, the researcher is not bound by individual beliefs or preferences of participants. If the metaphor expresses a general view, it is not problematic. However, it is admittedly difficult to distinguish values inherent in selecting the metaphor as such, and the researcher must be aware of the possibility of an inherent bias. Moreover, it raises questions about such things as member checking, a frequently employed strategy for ensuring rigour in the analysis. If participants disagree with some or all of the researcher’s approach to expression, whose view holds sway? The integrity and vulnerability of both researcher and participant come into play if disagreements arise around the findings. Balancing the harms and benefits of research is important in all phases of research from data collection to analysis and dissemination. Participants could be harmed if they feel the metaphor does not capture their experience or implicates beyond their experience. The question of how to deal with strong disagreements with the analysis, particularly if only one or a few participants disagree and the remainder agree with the metaphor as an accurate depiction of their views, must be considered prior to data collection and analysis. Although member checking could redress the power imbalance between participant and researcher, participant vulnerability may be amplified if there is inadequate discussion or planning for potential disagreements between participants and researchers when arts-based research strategies are employed.

Upon completion of the study under discussion, the major findings including the use of the river metaphor were shared with family participants and professionals clinically involved in this type of transition. This was not formal member-checking, but it allowed some feedback on the use of metaphor. Past participants and clinicians agreed that the metaphor had resonance and captured the transition experience. Despite the lack of evidence of harm, the ethics of using metaphor as a research tool attracted our attention. In particular, what if we translated our metaphor into another arts-based strategy, such as theatre or visual art? Art exposes. Aesthetics aim for vividness and increased understanding. The more blatant and public the art form, the greater is the potential for exposure and possibly distorted understanding of the concepts being expressed. When uncommon experiences combine with art, the risks to participant privacy and dignity increase. Researchers have a moral obligation to consider potential harms and not to sacrifice participants in the interests of vividness of expression.

The example presented here outlines ethical concerns arising from the use of metaphor in a text-based research analysis and presentation, but many of the same issues could emerge in research in which findings are presented using another arts-based form such as theatre. For example, if the researcher chooses to depict concepts or themes in the form of a play, the researcher must make every effort to ensure that the concept being expressed is not distorted in the interests of artistic licence. The question to be asked must be, ‘Will the participant agree that this representation accurately reflects his or her experience?’

Other ethical questions arise related to participant withdrawal. How and when can participants withdraw if their perspectives have already contributed to the development of the metaphor in its entirety? One can easily imagine a participant reviewing a metaphor in text-based research, or viewing a play ostensibly based in part on his or her experiences, and objecting to the form of representation. Although the researcher could argue that the putative distortion of ideas is necessary to get the point across, it is also possible that the participant might experience embarrassment or shame in having his or her experiences so represented. At what point can the participant ask that his or her contributions be withdrawn, and what obligation does the researcher have to prevent such harm? Although the researcher could put a limit on time on withdrawal, indicating that participants may withdraw only up to the time of data analysis, this does not obviate the concern about a negative response and harm caused by the representation of data. To ensure fully informed consent, the researcher will need to have discussed such possibilities before consent is signed.

A corresponding difficulty relates to the unexpected, or unplanned, use of artistic strategies. Blanket consent forms containing all possible analytical or dissemination strategies may not be possible as the metaphor emerges from the analysis, and cannot be predicted. Similarly, blanket consents may be considered invalid owing to their lack of specificity (Caulfield, 2007; Kosseim and Brady, 2008). Therefore, the concept of ‘process consent’ is vital (Cutcliffe and Ramcharan, 2002; Usher and Arthur, 1998). This approach offers researchers the opportunity to return to participants during the analysis, but this falls very close to member checking, and hence the original concerns of priority in disagreement remain intact. Values-based training of researchers might be beneficial, wherein participant dignity, integrity and vulnerability are prioritized and the potential implications of arts-based research on such values is discussed.

One final consideration for the researcher deserves mention. Ethical conduct requires that researchers communicate their findings once the research is complete. In writing up the research the use of metaphor can become a hurdle. Word limits and the novelty of metaphor in traditional health journals can lead journal reviewers to reject articles in which metaphor has a central place. In the case study presented here, the decision was made to present findings without the metaphorical reference (Manhas and Mitchell, 2012). The questions then arise as to whether, in departing from the metaphor in the manuscript, the authors failed to fully communicate the results of the qualitative analysis. Moreover, the novelty of the metaphor limited the appropriate journals in which to publish the results, and one might question whether that was an affront to the dignity of the research participants who had taken the time to participate.

Issues for research ethics boards

The kinds of concerns raised here speak not only to the obligations of the researcher, but also to obligations of the research ethics board (REB), or institutional review boards considering the proposal. One might question whether the REB has a duty to enquire about issues of distortion, values, and representation in the analysis and presentation of the research. There appears to be some agreement that REBs should examine methodology, because methodologically unsound research is inherently unethical (Canadian Institutes of Health Research et al., 2010). However, there seems to be little attention paid to the kinds of methodological questions raised here. Although the subjective nature of qualitative methodologies is often understood to fit well with the use of such an artistic approach to analysis and reporting, there is a question as to whether it is necessary to detail this in application for institutional ethics review. If it is not made clear at the outset, there could be a question as to whether participants have truly consented to such use of their data. In the study under discussion, for example, one might question whether veering from established approaches was a substantial or minor change to the proposed methodology. The former requires an approved amendment from the REB; the latter generally does not.

Ethics review aims to protect the dignity, integrity and vulnerability of participants, and REBs need to have a clear understanding of what exactly will be done if they are to protect participants from risk. In Canada, the TCPS2 acknowledges the somewhat fluid nature of qualitative research, and suggests that, because it is not always possible to anticipate the direction data collection and analysis will take, the researcher must be free to make minor changes through the course of the study (Canadian Institutes of Health Research, 2010). This is a strong position from the perspective of researchers, but it may be inadequate to protect the dignity and integrity of participants, particularly those who are vulnerable (as is often the case in health research). We are not arguing that REBs follow the ‘ethics creep’ towards broader, ongoing and invasive roles in the research process (van den Hoonaard, 2011). Current REBs, in Canada at least, lack the resources or support to extend beyond their current role at the proposal phase. Rather, the REB should be cognizant of the issues that might arise, and should ask questions of the researcher. Researchers have the obligation to be cognizant and to act, whereas the REBs have the obligation to be cognizant and to flag unaware researchers or their projects. To the extent that such problems can be anticipated, they should be addressed in the proposal, and participants informed during the consent process.

Conclusion

We have detailed some of the ethics considerations in using metaphor in data analysis and presentation in qualitative research. Issues related to participant dignity, respect and vulnerability led us to suggest that researchers should take these points into consideration in designing their research and seeking informed consent. It is important to have a plan in place for dealing with some of these issues. These should be detailed in the proposal and communicated to participants. REBs should, on their part, be prepared to ask questions if such details are lacking in the proposal.