Data collection

Qualitative studies encompass a variety of data collection methods, such as participatory and non-participatory observation, qualitative interviews, document analysis, video recording and audio recording. All of these methods have their own ethical issues. Regardless of the type of data collection, it is important to adhere to certain ethical principles, such as confidentiality, informed consent and safeguarding the research subjects' integrity. (See also Responsibility concerning the individual)


In qualitative interviews, it is important to safeguard the integrity of the interviewees both during the actual interview and afterwards, when the results are to be interpreted and presented. There is usually a requirement to anonymise recognisable data, and to treat sensitive information with discretion. In surveys that deal with sensitive issues, written consent is often required.

Questionnaire surveys

It is generally more difficult to safeguard the anonymity of research subjects in qualitative studies than in questionnaire surveys since the research subjects relate events they have experienced in their own words, and their stories are more recognisable to others. Questionnaires, however, have predefined questions and answer alternatives, which makes it more difficult to identify individual characteristics. Moreover, questionnaires are normally stored and interpreted at aggregate level, i.e. where it is only possible to see how a specific category of respondents is distributed in relation to a given variable, such as how many people with a higher education vote for the Socialist Left Party. Individual respondents cannot be identified when this method is used. It could therefore be said that the main ethical challenge in so-called quantitative research is its ability to provide static images of entire categories of people, such as when it is shown that the share of Somalis in employment or education is the lowest among all immigrant groups, but without any statistics to explain why this is the case or to show correlations in the lives of individuals. (See Quantitative methods and Statistics.)

Observational studies

The term observational study encompasses both participatory and non-participatory observation. In purely observational studies, i.e. non-participatory observation, the researcher observes other people without participating himself. An extreme example of this would be where the researcher listens to other people's conversations through a window. However, this method would make it difficult for the researcher to grasp the meaning behind what is being said and to form a clear understanding of what is going on. The danger of misinterpretation will clearly be present in such cases.

The observational role is illustrated in the film Kitchen Stories (2003, directed by Bent Hamer). According to directives from the research leadership, the researcher in the film is not supposed to talk to the man he is observing, or influence him in any other way. The film shows that it is impossible not to influence the subject when taking such an external observational role. On the contrary, the researcher's presence influenced the man to do completely different things to what he would normally do. Being under the watchful eye of a researcher making a note of everything that happens is such a stressful experience that it is inevitable that the subject behaves differently to how they would normally. The researcher effect may, therefore, be greater rather than less when a purely observational role is adopted. In addition, such an external observer role will be based on a problematic view of human nature when the research subjects transition into the alien "other", who is not met on equal terms through mutual communication. They become research subjects, without being allowed to have their own active role in relation to the researcher and research.

The only time we can envisage an external researcher role having no impact whatsoever on the research subjects is if the researcher is in a public area and behaves exactly like the others present, such as waiting for the bus like them, or sitting on a tram like them. However, if the researcher stares too much at the others, it can affect their behaviour. Knowing they are being observed will affect their sense of freedom and naturalness. A researcher who fits naturally into the surroundings therefore has a lesser effect on the behaviour of those present than someone who sits in a corner taking notes. If a researcher assumes the role of a voyeur – someone who sits in the bushes spying on others, or peering through their window without them noticing – it will not affect the interaction, but neither will he obtain very much meaningful data or a clear understanding of what is going on. Additionally, such a research role is compromised by the requirement for informed consent. In other words, the research subjects must be informed about the research, and they must give their consent to participating in the research project. The voyeur role is clearly incompatible with this requirement.

The most common method is a purely observational role, where the situations to be observed have a relatively formal and fixed structure, such as classroom lessons. In such situations, a completely isolated observer will be regarded as less disagreeable than when the most important social interaction takes place in a more spontaneous, informal manner. However, there are, of course, degrees of non-participation; an observer in a classroom, for instance, could supplement a purely observational role with some informal chat with students and teachers during breaks. Either way, the objects of study must be informed about the project, and that also applies to children.

In psychology research, it is relatively common to use non-participatory observation when children at play or a person or couple in a therapy session are being observed through a one-way mirror. In such cases, the objects of study are normally aware they are being observed and have consented to it, but there are also cases where observations are conducted without prior consent. Nowadays, research without informed consent is very much frowned upon, and observation without informed consent is not as common as previously. This is also related to the protection of research subjects' integrity. How can they feel their integrity is protected if they have not even had the opportunity to consent to being studied?

Participatory observation

Participatory observation is a term used more or less synonymously with field work, but relates more specifically to the principles that the method is based on: the researcher must simultaneously take part and observe. This, in itself, can create special ethical challenges. If the researcher gets too involved as a participant he may run the risk of letting the research and observation take a back seat. The other participants could regard it as a betrayal if the researcher who has gradually become as involved as them subsequently takes on a more distant and withdrawn researcher role. During lengthy field work, it is not uncommon for the researcher to form close bonds with some of the participants, and for friendships to develop. In such cases, analysis may prove difficult. In other words, close contact with certain individuals may be beneficial, as it gives researchers access and enables them to obtain continuous information, but careful consideration should be given in each case as to whether the close contact means the researcher is too close to use this person as a research subject.

As noted, it may be regarded as a breach of trust if everything a participant says and does is subject to analysis and interpretation. The fact that a confidential relationship of trust is developing therefore means that this person cannot be used as a research subject. It may therefore be better to allow some participants in the study to only provide information, and not act as a research subject to be quoted, described and analysed.

The researcher must also consider whether it is appropriate to let the relationship between researcher and respondent become a mutually binding friendship. A fusion of research-respondent relations and friendship can provide good data because the researcher can gain access to more sensitive information in this way, but it may also be problematic if repeating the information that is obtained in this way is regarded as a breach of trust. One way to solve this is by giving the person the opportunity to agree on what information can be used and, if appropriate, allow them to read what you have written before it is made public.

Hidden observation

So-called hidden observation, i.e. observational studies in which the researcher does not make it known that he is a researcher, or what the purpose of the study is, is considered by many to be unethical. Hidden observation can be an advantage when studying environments where it is difficult or impossible to gain entry, but many nevertheless consider this method to be ethically unacceptable, and therefore a method to be avoided (see for example Bulmer 1982). Hidden observation is completely at odds with the ethical principle of informed consent (as mentioned earlier in the description of the voyeur role). Defenders of the method often simultaneously reject the notion that this ethical principle should be the standard for qualitative research.

The infiltrator role is also different from a practical and psychological aspect; the researcher has to pretend to be someone that he is not. With infiltration, the researcher must play the role of a primary participant, and access into the field will therefore be dependent on being able to play the role of someone who wants to be part of the environment for the same reasons as everyone else. Notes must be taken in secret, and there is the constant risk of being found out. In criminal environments, the researcher may be exposed to danger and run the risk of reprisals if his true identity is revealed. One problem with infiltration when studying criminal environments or environments that express extreme and hateful attitudes, is that the researcher may be forced to participate in such actions and display such attitudes. It goes without saying that this would be an ethical dilemma. In this case, the researcher would be violating the principle of informed consent, in addition to committing a crime, as well as lying to respondents and establishing a fake relationship of trust with them. All in all, this role is completely unacceptable from an ethical perspective.

A more acceptable version of the same role is the researcher who analyses a phenomenon after taking part in it, as Aubert (1982) did in his study of the resistance movement during World War II.

Action research

In action research, the researcher tries to actively influence the field of study. However, changing the field in such an action-based research process is not normally done without participants accepting it. A common action research approach is where the change is effectuated through problem-solving cooperation between the researcher and the research subjects. In this way, the primary participants in the effectuated change are co-responsible, and thus the ethical dimension of the research process is safeguarded. More ethically questionable are any projects in which the researcher imposes a change on others that they do not want, for example, where, during the research, the researcher attempts to rehabilitate criminals or drug addicts. In order for such research to safeguard people's integrity, they themselves must be willing to participate in the change process in the research. Generally speaking, research should protect participants' integrity, freedom and co-determination. For research on potentially vulnerable groups, with which much of the action research is concerned, particular care needs to be taken, especially in cases where the research subjects are not able to fully understand what the research is about. This applies, for example, to research on children, persons who are mentally disabled, newly arrived immigrants or people who suffer from psychological disorders. Research on such vulnerable groups, and the assessment of the potential for change in the research, as well as the participants' degree of involvement and acceptance of this, will often be considered in relation to whether the research can provide findings that directly or indirectly, in the short or long term, can benefit society and culture (see NESH).

Hermeneutics, metaphors and a critical perspective

Today's hermeneutics were originally developed within philosophy and literary science and related to the interpretation of written texts (before this, hermeneutical methods were used in theology to interpret biblical texts). However, in the social sciences, the term has also been used about the reading of interview extracts, observations, and more generally - to quote an expression coined by anthropologist Clifford Geertz (1983) - "reading culture as text".

Hermeneutics in various qualitative studies also have their particular challenges when it comes to research ethics. In different variations of cultural analysis, either so-called double or triple hermeneutics are often used. The concepts were based on the idea that people interpret their reality, and that the social scientist's challenge is to interpret this pre-interpreted reality (Giddens 1976). The social scientist must, in other words, take into account that the social reality is already symbolically pre-structured (Schütz, 1967). This has an inner logic that is implicit in the myths and stories that the participants use to create their social context (cf. Habermas, 1984:107). This also means that the researcher's models differ from the models the research subjects have of their actions. The researcher's models are on a different analytical level; they are models of the research subjects' models. This is an ethical challenge in cases where the participants do not recognise the researcher's interpretations. Here it becomes a question of who "owns" the truth about a person's social world. If the researcher's interpretations are disseminated both to the group being interpreted and to outsiders, it can be a challenge when the research subjects disagree strongly with the interpretation. The researcher has a responsibility to avoid unfounded notions. When the researcher's interpretation is not consistent with the person's or the group's self-understanding, the demand for justification and documentation is strong. At the same time, research subjects cannot be allowed to act as censors of the interpretations; that would be in conflict with the requirement of freedom and independence of research. In some cases, the researcher's interpretations are so distanced and abstract that the research subjects do not fully understand the results. In order to reclaim the research, the researcher then needs to translate and explain the importance of the interpretations.

Metaphors are primarily a tool for identifying specific features of an object or event. Researchers often forget this and instead treat their metaphors as real things in the empirical world, writes Norman Denzin (1978:46). The metaphors then lose their explanatory value and become circular explanations. Using metaphors so that they can be a useful tool in illustrating meaning is therefore a difficult balancing act. Discussing and analysing the participants' own metaphor use is justifiable and important, but for researchers to develop their own metaphors as imagery for patterns in their own observations is far more difficult and must be done with much greater care. All too often, students end up trying to get their data to fit into the categories they have "invented", for example by dividing their respondents into three stylised types. A better approach would be to let the categories develop through an increasingly critical alternation between the reading of field or interview notes and reading of theory, and testing them out through constant re-evaluation of the material, where examples are also sought that do not fit the categories (so-called negative cases, cf. Becker 1970). The theoretical framework used as a basis for interpreting behaviour must in some way be connected to the reference frame of the participants' own primary interpretations. When an orderly and consistent argument is used that shows how the interpretations were derived, the use of interpretations that break with the research subjects' self-perception would also be ethically justifiable. The point is that the results should be credible, and the basis on which interpretations have been founded must also be understandable, including for the research subjects.

The interpretation of underlying, hidden interests and driving forces and critical interpretation can be regarded as interpretations of the third degree. They not only interpret the participants' interpretations, they are also critical of these, or search for hidden agendas and needs. Alvesson and Sköldberg (2010) refer to this as triple hermeneutics. Single hermeneutics relates to the participants' interpretation of themselves and their intersubjective/cultural reality. Double hermeneutics is what the interpreting social researcher engages in, by attempting to understand and develop knowledge of this reality. It is obvious that such critical interpretations, or interpretations of unconscious urges and needs, are more ethically challenging than the double hermeneutic interpretations.

The critical theory's triple hermeneutics encompasses the double hermeneutics and a third component, namely the researcher's critical interpretations of the structures and processes that, in different ways, affect the research subjects as well as the researcher's own way of seemingly freely interpreting their situation in relation to the topic of study. Critical interpretation explains unconscious processes, ideologies, power structures and other expressions of dominance, which means that certain interests are hidden at the expense of others in the forms of understanding that occur spontaneously (Alvesson and Sköldberg 2010). When the interpretation that people are ruled by a "false consciousness" is put forward in line with Marxist theory, this is questionable. An alternative way to interpret may be to demonstrate that the action strategies, identification methods and self-perceptions people have are understandable in light of the position they have in the sociomaterial structure, without thereby saying that they act on the basis of a false consciousness.

Instead of only interpreting underlying meanings in the broadest sense by "reading" actions as metaphors and narratives, interpretations can be based on a theory that the subjects under observation are acting on the basis of a false interest, or as a result of repressed urges and needs so that they are unable to see what the researcher sees. This starting point is also about views of human nature, since the researcher claims to have insight that is not shared by ordinary people. Instead of taking as the point of departure that an action can have an infinite number of meanings, the researcher then searches more for the action's actual underlying meaning. The researcher seeks hidden "truths" that underlie human actions. Marx's theory of economic interest, Sigmund Freud's notion of the subconscious and Friedrich Nietzsche's concept of the desire for power are all theoretical constructs that are based on the idea that such actual meanings exist. Ricoeur (1974, 1997) calls such interpretations reductive (restrictive) or hermeneutics of suspicion.

Psychoanalysis claims to be able to explain myths and symbols that use unconscious representations, such as distorted expression of the relationship between libidinal impulses (sex drive) and the repressive structure of the superego (which represents our conscience) (Ricoeur, 1997:318). Ricoeur (1981:34) is sceptical to this perspective and to Freud's approach to the feeling of guilt as an expression of neurosis or the repressive effect of cultural norms. According to Ricoeur, the conscious, pre-conscious and subconscious – as well as the ego, id and superego – are not actual occurrences, they are interpretations. Thus, according to Ricoeur, the subconscious is not something that actually exists, but rather a meaning we ascribe. As Ricoeur sees it, psychoanalysis does not operate with facts, but with people's life histories. The life histories can be viewed as a form of texts that it is the analyst's task to interpret meaningfully as a coherent story.

From a research ethics perspective, it is important in such cases to show composure in relation to the requirement to reveal the truth about the research subjects. Being subject to the research and observations of others can, as noted in NESH's Guidelines for research ethics in the social sciences, law and the humanities, in itself be degrading. This applies all the more when the researcher identifies urges and needs that the person himself does not acknowledge having. The individual life stories can be seen as symbols that can only be attributed meaning in light of the context of the text (the full life history). For a psychoanalyst, the goal might be to give the patient the self-insight to understand his own biography; the story of himself, as a history. From a research ethics perspective, such a therapeutic goal may, however, be questionable, especially if it is contrary to the person's integrity.

Interpretation of data/limits of interpretation

Regardless of the level of interpretation a researcher chooses, it is important that he is self-critical when examining his own interpretations, and that he endeavours to let a clear distinction be made in the text between his own interpretations and the participants' interpretations of their actions, as well as what are the participants' statements and what are the researcher's interpretations of these. In participatory observation, it is easy to confuse the respondents' interpretations with the researcher's interpretations. It is therefore particularly important to be aware of this. One way to prevent the distinction becoming blurred is to cite the participants' claims as quotes, and then interpret the claim. When the raw material is presented in detail, the reader can easily see why the researcher interprets it in the way he does and can also more easily distinguish between the researcher's interpreted version of what he has seen and heard, and what has been seen and heard irrespective of the researcher's interpretation. When interpreting the motives underlying a person's actions it is particularly important to proceed cautiously. Interpreting the motives of other people entails great uncertainty. It must be possible to clearly document and justify why particular actions are believed to have been based on specific motives, especially in cases where motives are ascribed to a person that differ from those claimed by the subject.