1. Introduction

The Universal Declaration on Bioethics and Human Rights recognises in its preamble that "scientific and technological developments should always seek to promote the welfare of individuals, families, groups or communities and humankind as a whole in the recognition of the dignity of the human person and universal respect for, and observance of, human rights and fundamental freedoms" (United Nations Educational, Scientific and Cultural Organization [^{UNESCO], 2005}, p. 1). It adds that "ethical reflection should be an integral part of the process of scientific and technological developments and that bioethics should play a predominant role in the choices that need to be made concerning issues arising from such developments" (^{UNESCO, 2005}, p. 1).

Article 1 of the Declaration describes its scope with respect to "the ethical issues related to medicine, life sciences and associated technologies as applied to human beings, taking into account their social, legal and environmental dimensions" (^{UNESCO, 2005}, p. 2). However, this reference does not appear to refer directly to the social sciences as such; they remain implicitly a dimension of other sciences.

The International Ethical Guidelines for Health-related Research Involving Humans of the ^CIOMS (2016) contemplate more extensively some social studies, such as observational ones, and include data collection techniques, such as interviews and surveys. Their scope is limited to biomedical or behavioral science from a more clinical or therapeutic paradigm. These guidelines appear to be the only normative-ethical instrument that incorporates aspects of social sciences, not only in the text itself, but also in the participation of social scientists in the drafting of the 2016 version (^{CIOMS, 2016}). However, the Declaration and the CIOMS guidelines, were not conceived specifically with the goals of the social sciences in mind. Yet the social sciences should be mentioned clearly in the Declaration and the CIOMS guidelines, as their status as sciences, through their contribution to scientific development and health-related research, is obvious.

Certain ethical conventions in the social sciences were established before the 1960s. The classic sociological studies of the Chicago School preserved anonymity through the use of pseudonyms, and they omitted geographical locations. Anthropologists also used pseudonyms for tribes and their members. Both disciplines addressed the ethics of covert observation and were able to arrive at a professional consensus and regulate themselves (^{Israel, 2015}), such that investigators had to meet specific ethical requirements in order to carry out their studies. ^{Israel, 2015} identifies the first code of ethics in the social sciences as that of the Society for Applied Anthropology, in 1948, which describes anthropologists' responsibility to avoid situations that affect the welfare of groups or individuals. The American Anthropological Association and the American Sociological Association then created their own codes outlining the obligation to protect research subjects from harm (^{American Anthropological Association [AAA], 2012}; ^{American Sociological Association [ASA], 2018}). There is a greater sensitivity to structural problems and broader social responsibilities in these latter documents than in the biomedical ethical guidelines of their times (^{Israel, 2015}). The associations avoided arguments about universality and demonstrated support for their codes through ratification processes.

The work of these associations was met with a certain amount of rejection of its nature and content, perhaps owing to a lack of ethical authority. This rejection led them to cede responsibility for ethical safeguards to research institutions and councils or IRBs, and these expressed the concerns that were outlined in the codes of biomedical ethics (^{Sieber & Tolich, 2013}; ^{Israel, 2015}). Bioethical principles and their practices have been applied not only to biomedical research; their requirements have also reached the fields of social science research, with varying degrees of scrutiny and success (^{Israel, 2015}), especially when the object of study is health and illness.

The question that arises from this situation is whether the requirements of the UNESCO Declaration and CIOMS guidelines or the declarations of social science associations, such as the British Sociological Association's 2017 declaration on ethical practices or the Social Research Association's ²⁰¹³ New Brunswick Declaration, have an influence on the way that social research is structured (^{British Sociological Association, 2017}). Do their specific requirements -in which risks and potential harms are usually interpreted differently in biomedical and social research-influence the way in which methods, procedures, and results are structured with respect to participants and communities, or with respect to sensibilities regarding the structural problems of individuals and their social responsibilities?

Article 4 of the UNESCO Declaration holds that "in applying and advancing scientific knowledge... direct and indirect benefits to patients, research participants and other affected individuals should be maximised and any possible harm to such individuals should be minimised" (^{UNESCO, 2005}, p. 3). It uses the term "harm or harmful effect" (in Spanish version) which refers to that which causes harm, which is detrimental, injurious, or adverse, and which has also been related to the concept of risk (^{Bosi, 2015}). The guideline 4: potential individual benefits and risks of research states that "...risks to participants are minimized and appropriately balanced in relation to the prospect of potential individual benefit and the social and scientific value of the research [...] "Risk is generally understood as an estimate of two factors: first, how likely it is that a participant will experience a physical, psychological, social or other harm; and second, the magnitude or significance of the harm. This understanding of risk implies that discomfort, inconvenience or burdens are harms of a very small magnitude that are almost certain to occur. This guide does not make a difference between risk and harmo(^{CIOMS, 2016}, p.9-10).

The application of the Declaration or CIOMS guideline to the question of reducing harm to participants is often disputed in the context of social research: it is believed that this type of research underestimates or minimises the harms, or fails to communicate them to participants (^{Pidgeon, Simmons, Sarre, Henwood, Smith, 2008}; ^{Sorokin et al., 2017}), that it requires fewer ethical safeguards, that its harmful effects cannot be anticipated, that it causes no physical effects on participants, or that these effects are different or simply non-existent (^{Santi, 2015}; ^{Sorokin et al., 2017}).

The difficulty in recognizing risks and harms in social research has been generated by, among other factors, the use of these terms without specific definitions and the failure to identify them during research. It is known when studies are planned that risks appear from the beginning, and that harm can emerge in the course of research or at the end, when researchers consider whether there were harmful effects and evaluate the type of harm (^{Santi, 2015}; ^{Sorokin et al., 2017}; ^{Aarons, 2017}).

The "risk estimation is highly inaccurate, whether based on some kind of past performance or on a future projection". Assessing the harm resulting from research involves understanding when and under what circumstances a participant may be harmed, and the justification of restrictions on personal freedoms to prevent harm to third parties, as well as taking action against harms caused (^{Santi, 2015}). For this reason, such assessment does not necessarily adhere to the biomedical concepts of calculation and predictability of risks (^{Sieber & Tolich, 2013}, p. 22; ^{Bosi, 2015}; ^{Sorokin et al., 2017}).

The bioethics of research on human subjects, including social research, considers harm in the following terms: a) physical harm, which involves pain, suffering, or injury; b) psychological harm or affective , which can manifest itself in the form of depression, anxiety, guilt, or stress; c) social harm, which exposes participants to harm in their relationships with other persons and is related to private aspects of human behavior, such as stigmatization for a particular illness or behavior outside of common moral norms; d) economic harm, in the form of monetary loss or loss of employment; e) legal harm, related to problems within the justice system for acts committed in the past or present; f) environmental harm and g) cultural harm, related to community (^{Santi, 2015}; ^{Sorokin et al., 2017}; ^{Aarons, 2017}). Harms in social sciences are related to psychological distress, social disadvantage, cultural values of a community and invasion of privacy more than to physical injury, although the latter is considered (^{Israel, 2015}; ^{British Sociological Association, 2017}; ^{Sorokin et al., 2017}; ^{Villaroel, 2020}). That is, harms that are "major, severe, or catastrophic are not common in social research, although they are possible" (Santi, 2015, p. 18).

Risk can be understood as the possibility or probability of occurrence of an undesirable event or negative result (^{Echemendía, 2011}; ^{Sieber & Tolich, 2013}; ^{Santi, 2015}; ^{Aarons, 2017}), and can be considered from different points of view: 1) quantifiable, based on probabilities of occurrence, whose mathematical expression varies from 0 (impossible) to 1 (certain); or 2) non-quantifiable, admitting a high degree of uncertainty (^{Bosi, 2015}). Risk depends on a subjective evaluation of the magnitude and severity of possible consequences or repercussions (^{Israel, 2015}; Santi, 2015; Aarons, 2017). For example, Decision Theory postulates that people's decisions are permeated by their assessment of risks and when these seem familiar to them their level of concern decreases. Conversely, they view unusual or unknown risk as deserving greater attention, given the potential danger involved. The background to this cognitive bias remains the subjective assessment and magnitude of risk. (^{Sieber, Tolich, 2013}; ^{Israel, 2015}; ^{Santi, 2015}; ^{Aarons, 2017}).

The most common risks in social research are during data gathering, in the management of relationships within communities, and in the practical aspects of getting to research locations (^{Israel, 2015}). When researchers initiate direct contact with study participants, their mere presence in the field generates expectations or ideas that may differ from or exceed the goals of the study. When studies involve sensitive topics, such as experiences of violence, abortion, suicide, there can be emotional risks, such as possible revictimization in the reliving of a traumatic event (^{Pidgeon et al., 2008}; ^{Israel, 2015}; ^{Sorokin et al., 2017}; ^{Villaroel, 2020}). For social scientists, risk evaluations should consider these potential impacts or harms that their studies may have on participants, with attention to subjective, ethical, moral, and cultural factors that influence individuals' decisions, and how these are employed in the situations of uncertainty, unpredictability, and intersubjectivity that are inherent in social processes. Assessments should include a reflexive component to identify different types of harm, and they should avoid quantifiable and predictive criteria to establish the potential level or degree of harm (^{Bosi, 2015}; ^{Israel, 2015}; ^{Sorokin et al., 2017}).

Although the CIOMS guideline incorporates aspects of the social sciences, the concepts of risk and harm are not clearly distinguished; risk is used as an indication of the probability and magnitude that a harmful event may occur. Article 4 of the UNESCO Declaration, uses the term harm or harmful effect, which describes what causes damage, harm or is pernicious, as well as its association to the concept of risk (^{Bosi, 2015}; ^{UNESCO, 2005}); this precision of the concept is more appropriate in the field of social sciences and is what generated the selection of the Declaration for this work.

This study focuses on an approach to the praxis of social research that analyzes the harmful effects that may arise in this context, and describes how Article 4 of the UNESCO Declaration, Benefit and Harm, applies to social research on health. The objectives of the study were: 1) to understand the harmful effects in social research; and 2) to describe the application of Article 4 to harm reduction in this field of research.

2. Methods

This was a descriptive, cross-sectional study in the form of a nationwide survey of social scientists in Mexico who had conducted research in the previous three years on topics of mental health related to violence, addictions, or suicide. The selection of topics was based on their current importance on public health and in the fact that this type of problems can be studied by the biomedical and the social sciences. Participants were from different educational and health-related institutions. The survey was carried out from February through May 2014.

2.1. Participants

Contact with participants was carried out in two phases:

Phase 1. Identification of social scientists. An exhaustive internet search was conducted for social scientists who met the inclusion criteria. The search process consisted of: 1) a search for researchers on the websites of institutions conducting social research; 2) selection of those researchers who met the inclusion criteria; and 3) collection of public information about the social researchers.

Phase 2. Selection of participants. Participants were selected as a convenience sample. The inclusion criteria were: 1) social researchers from public academic or research institutions who had conducted social research on mental health in the period 2010-13; 2) affiliation with institutions recognised by the National Council on Science and Technology (CONACyT) and registered with the National Register of Scientific and Technological Companies and Institutions (RENIECyT); and 3) affiliation with institutions under the Coordinating Commission of National Health Institutes and Specialised Hospitals (CCINSHAE) of the Mexican Secretary of Health (^{Secretaría de Salubridad y Asistencia [SSA], 1984}).

The search found 136 individuals conducting social research on the topics and within the period specified. All were sent an email inviting them to participate in the survey. Only 62 (45.6%) responded; the 74 non-responses (54.4%) may have resulted from outdated email addresses or from the invitation being diverted to a spam folder. Nine (14.5%) of the researchers responding declined the invitation, and another 6 (9.7%) did not complete at least one section of the questionnaire and were excluded from the analysis. Of the total of 47 participants whose responses were analyzed, 22 (35.5% of those responding to the invitation; 47% of the total analyzed) responded to one section of the questionnaire and 25 (40.3% of those responding to the invitation; 53.2% of the total analyzed) responded in full. A description of the sample is provided in Table 1.

Table 1. Sample characteristics. 

2.5. Data Analysis

The analysis included only information from responses to the open-ended questions in the survey; this data was condensed into a single text file containing 1,955 responses. A thematic analysis (^{Howitt, 2010}), which identifies broad themes that summarise the content of data, was performed. Although this is a type of analysis used in qualitative research, it is less demanding than other techniques; because it is not closely associated with a theory, it is appropriate as a descriptive tool.

A descriptive coding of the data was performed, with each word, phrase, or statement referring to a topic assigned a code. During this process some codes that proved inadequate were subdivided or corrected, and others with overlapping meanings were combined. The process produced 2,989 codes for the 1,955 responses. The most frequently appearing codes (those that appeared more than four times) were then combined, producing 279 distinct codes.

The next stage of the analysis was the identification of the theme. Constructs were examined that connected a substantial number of codes. These were grouped together in terms of similarities and common characteristics, which allowed for an evaluation of whether the themes reflected relationships or differences between codes. The themes constructed captured the general meaning of the descriptive coding. Throughout the process of constructing themes, the relevance of each was reviewed with respect to the data as a whole and also between new themes with the textual responses to the questionnaire. The grouping together of the codes and the development of themes were carried out by members of the research team in order to compare information and come to agreement.

The final stage of the thematic analysis was a search for literature concerning social theory and other texts related to ethics in social research in order to support the interpretation of the codes and the themes to which they could belong. To complete the analysis, the responses to the questionnaire were again reviewed, this time in light of the themes that had been defined, in order to observe the relevance of each theme to the participants' responses. The importance of each theme (see Table 2) was based on its relevance to the purpose of the study.

Table 2. Thematic analysis process. 

3. Results

The thematic analysis yielded nine themes associated with ethical aspects of social research on mental health: 1) obtaining and managing data; 2) ethical implications of social research; 3) causes and effects in social research; 4) expected impacts of social research; 5) research skills and abilities; 6) actions in social research; 7) implications of the research setting; 8) use of incentives; and 9) regulation.

This study considered only one part of theme 3, causes and effects in social research. Participants defined the theme as follows: Social research is characterised by the use of a series of techniques or actions that can promote or cause possible risks to study participants by researchers who, in some cases, necessarily have a direct or indirect effect on them. Based on the data, these were divided into three types of causes and effects: 1) causes based on the actions and decisions of the researcher; 2) effects on participants; and 3) consequences of the research. The latter, however, appears to refer to a possible redress of harms, which could be interpreted as the application of Article 4 of the UNESCO Declaration. Results for each of these three types were as follows:

4. Conclusions

The possible risks in social research that survey respondents were able to identify were the following: inappropriate use of and manipulation of data, generation of false expectations in participants, and the invasion of their spaces and intimacy. The harms identified were the revealing of participants' identities, the violation of their rights, and the generation of stigma and prejudice. These responses also showed other types of harmful effects, related to researchers' individual interests or those of their institutions. These negative consequences of social research were attributed to the decisions and actions of researchers and also to certain institutional demands. With respect to the minimization of harms stipulated in Article 4 of the Universal Declaration on Bioethics and Human Rights, respondents mentioned actions to redress or avoid harm to participants and researchers, including suspending interventions, providing emotional containment, and referring participants to appropriate support according to their needs.

The findings suggest various points for discussion. The first is that the survey respondents did not minimise or underestimate the risks and harms in social research; they are even clear about the different terms (there is risk at the beginning of a study and harm at the end). Second, they identified the harms as structural problems or as part of their social responsibilities, implicit in the decisions of researchers or the institutions they represent (^{Israel, 2015}; ^{Sorokin et al., 2017}; ^{Villaroel, 2020}). Third, they associated harmful effects with ethical, cultural, and subjective factors in the uncertain situations and unforeseeable procedural circumstances that are inherent to social research (^{Sorokin et al., 2017}; ^{Villaroel, 2020}). Fourth, Article 4 of the Declaration has an important influence on social researchers' practices in avoiding harmful effects.

According to the respondents, the harmful effects of social research depend on the cultural and social contexts in which the studies are carried out, and these contexts explain the phenomenon of research on over studied populations or those in vulnerable situations. In their responses concerning the minimization of harm to participants there was a close relationship between the ideas of public apologies, legal complaints, and the discrediting of researchers who commit harmful acts. In this sense, it seems that the harm to a researcher's reputation implies more a justly negative consequence for the researcher than a compensation for harm to the participant. When it is a matter of making decisions in the face of an undesirable consequence or risk in order to reduce the harmful effects as much as possible, this is usually based on legal, academic-political, and moral interests- for example, in apologizing.

The Declaration states that "in applying and advancing scientific knowledge", possible harmful effects to participants and others involved in research activities should be reduced to the extent possible. The survey respondents described the reduction of these effects as a question solely of the decisions of researchers, not necessarily a deliberate or reasoned and reasonable process (^{Villaroel, 2020}). For example, deciding to go beyond what is planned or studied is a decision that may imply various courses of action, but it is in a way a negative or risky choice because it can end up affecting the people being studied.

The Declaration is an attempt to provide new approaches to social responsibility in order to guarantee that scientific and technical progress contributes to justice and equity, and that scientific and technological advances are shared by all. If harmful effects in social research are not reduced, if researchers do not assume responsibility for ethical precautions in their studies, it will be difficult to develop an appropriate, fair, and equitable social environment in which participants are beneficiaries of scientific advances.

Harmful effects can be understood as inherent in any type of research; the problem is when these are caused by improper decisions or actions of researchers, whether intentionally or accidentally. Confession, minimization, and compensation of harm to participants is not sufficient; neither is an apology- the minimum element necessary for treating them in a dignified manner. A focus on discrediting, punishing, or firing researchers is even less satisfactory, as these are forms of moral lynching. The process should be oriented toward planning and foresight, with consideration of the ethical dilemmas that could arise in order to guarantee that harmful situations do not occur.

It is imperative to reorient the fundamental objective of social research, whose findings ought to generate a change in the quality of life of people who are unjustly treated or who are the mere enactors of social phenomena. It is essential to encourage research based on achieving the greatest benefit and the greatest reduction in harmful consequences. according to the characteristics and needs of the context in which it is carried out.

Paragraph (e) of Article 2 of the Declaration cites as one of its purposes the fostering of "multidisciplinary and pluralistic dialogue about bioethical issues between all stakeholders and within society as a whole" (^{UNESCO, 2005}, p. 2). This purpose could include the social sciences, given that the Declaration is an attempt to universalise an ethical framework in scientific endeavor that is increasingly breaking down borders. Although it was not expressly intended to consider the theoretical and methodological requirements of the social sciences, it should be useful in these fields, since it can suggest links between the regulations it expresses and the ethical practices required of social scientists.

One link between the Declaration and social research is found in the use of the term "harmful" instead of "risks" or "damages", which usually relate to the biomedical context and create resistance to the idea that such research may cause harm. Social scientists have also proposed harm as an alternative concept that refers less to a characteristic than to a quality. This language of qualities preserves an epistemological coherence with the social approach (^{Bosi, 2015}).

With respect to the adoption of Article 4 of the Declaration, the reduction of harmful effects in the social sciences as a condition of ethical action and practice reflects a commitment and moral responsibility toward others, which contributes to the consideration and construction of ethical values like freedom, solidarity, and social justice, through ethical dialogue, that allow for a symmetric relationship where all participants are recognised as equals, regardless of the needs and limitations they face because of changes in their lives.

Although the egalitarian position in itself can either reduce or exacerbate imbalances of power (^{Tealdi, 2008}), it is necessary, because it does not refer to the exercise of power, but to dialogue, which can encompass the interests of researchers and participants in an enterprise such as social research.

The observation of and respect for ethical considerations in social research is a topic requiring further study, since the intersubjective interaction of researcher and subject in particular settings allows us to under the harms and benefits of such research.

Ethics approval statement

The Research Ethics Committee of the Ramón de la Fuente Muñiz National Institute of Psychiatry (INPRFM).