Our team of enumerators recently tried to interview randomly selected households in rich suburbs in Johannesburg. They couldn’t get anywhere near them. Security guards stopped them; in some cases our enumerators (who are black) were told that black people aren’t allowed to ring the doorbell. Interestingly enough the domestic workers and security guards of the rich were more than happy to participate in the research.

While we did not manage to collect any data, the process was not a complete failure as far as we are concerned. We learned first-hand about the practical difficulties of trying to enumerate the rich face to face, door to door. That is critical learning, not only for the business of conducting primary research, but also for the business of analysing survey data that purports to capture the views of the rich. We now regard survey data on the rich with even more scepticism than we do other data.

It also did lead us to wonder how else we could get data on this segment of the market that is so intent on escaping enumeration. Given the willingness of domestic workers to participate in the research it got us thinking; we don’t actually need to speak to the madam to find out what brand of dishwashing liquid her domestic worker uses to wash the dishes, or what brand of washing machine she uses, or even when the madam bought it. We could just ask the domestic worker. In fact when you think about it, there is not much that happens in a typical upper income suburban home in South Africa that the domestic worker doesn’t know about. Chances are she makes the bed and washes the underwear.

We shopped the idea around socially within our network of madams. Most of them were not too worried about the prospect of their domestic workers answering questions about shopping and consumption patterns of their employers. But then there were a few who were horrified. Not just horrified but really HORRIFIED. The relationship, they felt, is tenuous at best. It is riddled with guilt on the part of the madam that she has so much and the domestic so little. Some were concerned that the research would highlight inequality, and throw it up for scrutiny in the face of the maid, as if she hadn’t noticed.

We therefore informally tested the idea with a member of the ethics committee at one of the universities. Her response was even more outraged than that of the most horrified madam. It violates a key principle of research ethics, namely the autonomy of the subject (in this case the madam) and the requirement for informed consent. To quote from a paper on medical research ethics:

“Autonomy refers to the right of an individual to determine what activities they will or will not participate in. Implicitly, full autonomy requires that an individual be able to understand what they are being asked to do, make a reasoned judgment about the effect participation will have on them, and make a choice to participate free from coercive influence. The cornerstone of protecting autonomy is the informed consent process, whereby an investigator provides a potential research participant with full disclosure about the nature of the study, the risks, benefits and alternatives, and an extended opportunity to ask questions before deciding whether or not to participate”1

We scrapped the idea and now we have no unique data about the rich. While it is frustrating in business terms, to be truthful, it is not a great tragedy in terms of furthering our knowledge. The most interesting thing about the rich is that they have a lot of money, and we knew that before we started.

The poor, on the other hand, are way more interesting. The consumption decisions they make, the risks they face, the trade-offs between today and tomorrow; these are fascinating decisions that we would do well to understand better.

Traditionally, research on the poor was strictly the domain of academia, after all it is not easy to monetise insights on a consumer segment that has no money. But this has started to change. Increasingly companies are interested in the poor as a market. Some may have been reluctantly pushed to develop this interest by the State, or encouraged by donors who wish to see greater inclusion of the poor in market economies. Nevertheless, there is more research on the poor that is being conducted by commercial entities such as our own, for commercial purposes.

Once again, we must comply with the requirements of standard ethical research practises. As noted above this includes autonomy; the poor must obviously give informed consent to participate in the study. It also includes a range of other requirements such as disclosure regarding the use of data, the identity of the sponsoring entity and the research purpose, commitments to protect the privacy of the data and restrictions on the use of data for sales purposes[2]. The principles are sound. Whether they are applied is another matter.

Even assuming they are, the ethics of research takes on a completely different hue when it comes to the poor. Firstly there is the issue of informed consent in the context of the always uncomfortable relationship between those who have some money and those who have none. Then there is the issue of harming the interviewee. As the concerned madam pointed out, sometimes by asking questions about the circumstances of their lives researchers cause distress to respondents whose lives are difficult. Then of course there is the issue of remuneration. Is it ethical to ask the poor questions and then use that data for commercial gain without sharing that gain with the respondent? Offering participation incentives is common practise when doing research in higher income, more time-constrained segments of the market. But these are often not demanded by poorer respondents. Further paying respondents to participate in some forms of research might undermine the research objective.

But perhaps the most troubling issue relates to the ethical obligations that arise from an interaction between the researcher who typically has access to resources and a vulnerable, needy respondent.

Let me clarify with an example: You are doing some research on expenditure and savings patterns of very poor people. You interview a family who is hungry. You ask them questions about food in the house and there is none. Or another scenario; you are doing work in a household where a child is sick and there is no money for medical care. Is it ethical to walk away without helping? What if the research requires you to come back and visit the same households again and again, with the intention of exploring how households cope with financial shocks, how they save and consume. Naturally, you will build a relationship with the household. Does this change the ethical playing field?

The question is simply: what does a good, ethical market researcher do in the face of human need?

The spectrum is bounded by two extreme positions. The first is that research on the poor is like wildlife photography. You are there to observe and record. Life is cruel, but you can’t intervene. If you do, you interfere and introduce noise into your findings, and your research is compromised. At the end of the day, a hungry household, or a sick or even dead child is a small price to pay for research that will save hundreds.

At the other extreme some might say that before you are a researcher, you are a human being. You cannot stand idly by and record the tragedy of a life poorly lived in the interests of research.  Your research subject is not an object. He or she is a person. Just like you. Help when you can. Even if it means you have to purge the data. Better that than having to purge your soul.

Of course the more compassionate approach has its own shortcomings. Researchers cannot help every respondent and the obligation to do so would mean that the research never gets done.

A third approach might be to structure the research so that the well-resourced researcher doing the research does not come into contact with the poor respondent. The researcher could train and pay community-based enumerators to gather data. In many cases this not only circumvents confrontation with any ethical issues, it creates a benefit if not for the respondent, then at least for the community (via the enumerator) and may even improve the quality of the research.

Of course this work-around doesn’t remove the complex ethical questions. Clearly they remain. There may simply be no clear-cut resolution – just as there is none to guide the many day-to-day interactions we have with others who have so little. But that does not mean the questions should be ignored.

By Illana Melzer

1 See Research Ethics by ELizbeth A. Adams, PhD, CIP (2013) with contributions by Timothy Callahan, PhD (1998), ETHICS IN MEDICINE   University of Washington School of Medicine, Http://depts.washington.edu/bioethx/topics/resrch.html

2 A good summary of these is the US-based Market Research Association’s Code of Marketing Research Standards. According to the MRA’s website, “the Marketing Research Association is the leading and largest U.S. association of the opinion and marketing research profession, which delivers insights and strategies to help guide the decisions of companies providing products and services to consumers and businesses.” See its code of conduct here http://www.marketingresearch.org/issues-policies/mra-code-marketing-research-standards