Illustration by Margherita Miani

’Are you biased?’: the complexity of researching & advising on inclusion in the public eye

Rosalie Post


In this paper, I ponder a research project that I was a part of, during my employ­ment at a research consultancy. I want to use it as a case study, to raise some ethical questions I have no definite answers to: how do we do research on diversity & inclusion, particularly for public policy, and how do we do it in an ethical way? And what are the best/right policy choices? When it comes to the positionality of a research consultant and policy advisor in housing (this was my job title), how could I have ever known for sure if my advice was any good?

The paper dives into a project I did a few years ago where we were asked to do research on the possibility of discrimination among a student population who choose their own house mates from the candidates. The housing is cheap and good quality and the waitlist for it is long. The case highlights the fickle and emotional nature of the topic of housing, with insiders and outsiders and a multitude of perspectives. I would like to even ask: Is dividing a scarce resource ever ethical? And if we conclude that it’s not, and decisions still have to be made in the public sphere, what do we do then?

AnthroArt Podcast

Rosalie Post


Rosalie is a design anthropologist, driven by the belief that the world needs to be a better place and that we need better decision making to make it happen, and that it starts through a better understanding of how humans really work. With extensive consulting and strategy experience in housing and public policy, she transitioned into training social scientists to find solutions for wicked problems. She is the co-founder of Namla, owner of Berkenblad Antropologie, and one of the six partners in the AnthroArt project. She is also working on a book with a co-author on what business could learn from anthropology for future-forward decision making. 

Margherita Miani


Originally from Udine, Italy, I completed my master in Architecture at the University of Ferrara in 2016. Immediately after I moved abroad, first to Amsterdam and later Berlin, where I still am based.

In the past 6 years I have been working as a landscape architect for several offices, completing various tasks, ranging from large scale urban design projects to more detailed executive planning. Together with the design of public spaces, I always found interesting to explore how an architect can communicate their own idea to the public, especially when it comes to projects that involve participation process.

Drawing has been one of my biggest passion since childhood and in the past few years I have been looking for opportunities to make it a more important part of my professional life. I use my illustrations as a tool to express my inner world, my feelings and thoughts. I also get a big inspiration from observing people around me and I am especially interested in their practices of self care and what makes them feel closer to themself.

Find Margherita on Instagram: @marghe.illustrates.things


Imagine a story that happened a few years ago. One morning, the local newspapers in a mid-size Dutch city all had the same headline: City council accuses HappyStudent housing association of discrimination of prospective tenants. The articles underneath the headlines are based on a city council meeting that happened the night before, where several city council members heavily implied, but never actually said, that HappyStudents are racist, ableist, and/or classist. The case they are talking about, involves an anonymous student, who contacted city council, to report that they had been denied a house. From context, I and others have inferred that that student was a person of colour.

At HappyStudent headquarters, a discussion ensued. As one of the big three student housing associations in the Netherlands, present in many cities with universities and applied universities, and as a semi-government organisation, HappyStudent naturally has proper policies in place, explicitly stated on their website and all internal documentation, that they do not ever want to discriminate. Their system is based on wait-time, meaning that irrespective of who you are, if you wait enough time, you will be able to reply to room adverts in shared student housing, and be invited for viewings. The tricky thing is how the choice is made: in the Netherlands, it is common for a shared (student) house to do what is known as ‘hospiteeravond’[1]; a house invites several prospect house mates at the same time, socializes with them for an hour or so, and afterwards the house mates vote which of the ‘hospiteerders’ will fit in the house the best, and pick that person.

Internally, the organisation HappyStudent was divided on the issue of whether their current system could lead to discrimination or not. Sure, there was the person who called city council, but did they not get picked at viewings because of the colour of their skin? Maybe they were very shy, or had seemed arrogant during the ‘hospiteeravond’, or very nervous. About half of the organisation said, that the right to choose who you live with (so you can feel comfortable in your home), was more important than placing every student fast. The other half of the organisation felt that the potential of discrimination was a grave and big thing and they should switch to a system of placing students in shared houses at random. Their internal division is an example of a big and complex juxtaposition that plays out all over the world: the insiders right of feeling comfortable where you are (think also NIMBY, gated communities, etc.) and the universal right of every person to housing (among other places, ratified in the EU Declaration of Human Rights). It is one of the main unsolvable tensions that housing policy must again and again renegotiate. So their internal division was very logical. And they called us to commission a research to figure out whether discrimination was really going on.


So who is this ‘us’ they called? A few years ago, I was working in my previous job as a research consultant and policy advisor at an independent small social science research firm in the Netherlands. We advised governments and semi-government institutions about housing policy, either by providing statistical analysis, by giving policy advice with specialist expertise on laws and national government policies, or by doing qualitative research (less often than the other two). We had (and they have) an excellent reputation, and get called for the complicated questions. So, we were called in this case.

Now I think it is important to give a brief context about the system of social housing in the Netherlands, as it is so specific from country to country and it matters for this story. In the Netherlands, out of the roughly 8 million houses (housing units, including apartments etc.) that exist in the country, 3.4 million are rentals. Out of those, 2.3 million are owned by social housing associations, so 29% of all houses are social rent, and 68% of all rentals are social housing. Social housing is, as the name implies, rent controlled; the associations were privatized (in the nineties), and so fall into a category that I have come to call semi-government: bound by many laws and rules about who to house, when, at what price, etc., and not authorized to make a profit, but the government also backs their loans making it possible for them to borrow capital at much lower interest rates than a private counterpart.[2]

For students, there is rent-controlled student housing, in the form of rooms in shared houses and, if you are lucky and an older student, in self-sufficient studio’s, at affordable rates. Students sign temporary rent contracts that stipulate that they must leave their student housing 6 months after they graduate. Although they are not happy with it when that happens to them, it keeps the student housing market available for younger students. Most Dutch students these days find a rent-controlled room halfway through their second year, or at the start of their third.

One last thing is important to mention about the context of this piece: HappyStudent, the housing association, had a back-up rule in place: If you ‘hospiteer’, you interview for a house 20 times in 3 months, and you have not been chosen by any house, you will be assigned a room, in a random shared house.

Researching discrimination

HappyStudent considered this accusation of discrimination a big deal. Not only would it be illegal for them to discriminate, most if not all of the people who work there see their work as societal, important and selfless work, and they care about doing the best job they can. So they asked us to design a thorough research, and to discuss with them what we could do and for what sums of money. We, the project team consisting of a senior advisor, myself and two other mediors representing qualitative, quantitative and questionnaire expertise respectively, and a junior researcher, had a few meetings with them discussing different research techniques and what we estimated would be gained from employing those. But we also spent quite some time internally discussing how to research this as it was a challenge we had not tackled before.

We had experience of researching discrimination with rental brokering services; we would do ‘secret shopper research’ for that (a volume of emails from different email addresses/names to test who gets invited for viewings). But here, secret shopper tactics weren’t an ethical option; we were not willing to recruit students and select them for discrimination-prone aspects, to go to house viewings for us. Apart from it not feeling ethical, it would not have lead to reliable data, for it would still not answer whether they were not chosen because of their characteristic or because they didn’t do as well as another in the interview. Maybe another person present had the same hobby as one of the house mates and was chosen for that reason, and the characteristic would have had nothing to do with it.

Approaching people and interviewing them directly about their own bias would not have worked; if you ask any person on earth: ‘do you discriminate?’ they would most likely say no, and particularly most persons under 25 are much too aware of peer pressure and fitting in to give an honest answer, if they do discriminate. This is testing for someone’s awareness of socially acceptable behaviour, not testing their attitudes.

We also did not have the possibility to do lengthy participant observation. Not only were we not students, and would have stuck out in the environments we would want to observe, we also did not want to present an exorbitantly expensive research proposal to a social housing association, knowing that they spend money on us that they can’t spend on fixing and building houses.

What we ended up presenting to them, was a combination of statistical analysis and comparison with previous research (cheap, fast, using resources that are out there), quick questionnaires to a large audience of students (cheap, fast) and group interviews with student houses (expensive, but for deep insight). For the group interviews, I designed a number of roundabout ways of asking about bias, and watching how the conversation in the group would unfold. More on the interviews later.

The reason I am describing how hard it was to find a good way to research this policy question, is because it touches on a larger point in applied anthropology/anthropology in the public sector: there is a time and money constraint (and for good reason, not wasting money meant for the public good), and yet, the outcomes of your research must lead to policy advice, and can’t just be observations without conclusions. The anthropologist in me would have loved to participant-observe in the student housing blocks for weeks, and wait until I overheard conversations about bias which would have inevitably occurred at some point. But in the context I was in, my job was to design a qualitative research component about a very sensitive topic, that is fast (their question was urgent), indirect (you can’t risk insulting a tenant when employed by their landlord), concrete (leading to policy advice) and efficient (not wasting money or time). Whether or not I proposed the best research technique possible, I don’t know; I had two days to come up with it.

The research we ended up doing

What we ended up doing, is designing a quick questionnaire, accessible through a QR code, so as to cast a wide net. We stood on the street and at the entrance of student events, and university buildings, for a few days, with a flyer that said: ‘HappyStudent wishes to improve its services and understanding of your needs, please fill out this questionnaire’. The questionnaire would take most participants around 5 minutes to complete, yet in the back end it had many different paths; it asked if you were a new student this academic year or you were already here for some time, and then to new students it asked about their expectations of finding a house or experiencing difficulty in interviewing for a house, while older students were asked about if they had a house and if they had had trouble finding it. It measured in this way both perceptions of younger students (that could colour their experiences while interviewing for houses) and recorded actual experiences of being chosen or not chosen. If someone answered that they had not been chosen many times, we asked why they thought that had happened to them in a ‘type an open answer’ question. We also asked if they were willing to share any characteristic they had that they thought could label them as ‘different’.[3] Only at the end of the questionnaire one would know what the research was about. HappyStudent incentivised participation by giving away a year’s worth of subscription to a bike-rental service, that most students in the Netherlands already use. 600 students filled out the questionnaire.

We also did 15 group interviews with student houses. HappyStudent kindly incentivised participation in these as well: for every house, we brought a crate of beer, and ordered everybody in the house a pizza. Some houses have 16 rooms to a corridor with a shared kitchen, other houses have three bedrooms, so interview groups differed in size. We tried to create a loose atmosphere in every one, construed it as casual conversation. In the first part, we asked them about how long every house mate had been living there, when their last ‘hospiteeravond’ was and what it was like. This usually led to stories about funny misunderstandings, or complaining, or other topics we could play into to bond with them, and gave us an understanding of the dynamics in the house. In the second half of the interview – which was timed to match the arrival of the pizza – we gave them all a laminated paper with AI generated portraits, and made up names, ages, study directions and hobbies or preferences. We asked them: ‘If these 16 people would show up to your next hospiteeravond, who would you want to start a conversation with first?’ and after many other questions we asked them lastly, ‘Who of these 16 would you think has a small chance of being chosen?’ We ended the interviews by telling them these were AI generated pictures.

The results surprised me: there seemed to be a house for everyone. We visited a house where 15 people yelled over each other loudly for 2 hours about partying. We were in a house where everyone had a form of disability or chronic illness and all wanted to be asleep by 10pm. We were in houses where everyone was white, but also where almost no one was; in houses where no one or everyone identified as queer or religious. At the same time, no house tolerated everyone; partygoers were fine with gender, background and race, but didn’t want early sleepers or shy people. Christians were okay with every race, gender, personality and sleep-preference as long as they were also Christian. Queer people were afraid to not be respected by people from many cultural or religious backgrounds, so apprehensive about living with them.[4] So everyone had a bias, and at the same time, I could not identify a group who had no chance of fitting in anywhere. Many students told me in the interviews they ended up in the right house for them in their second or third year, sometimes at their second or third try. They recounted feeling rejected when they were interviewing for other houses, until they got that right place for them, and then they were glad to have found it and understood the point of hospiteren.

I also spoke to houses where someone had been placed through the back-up rule, and the results weren’t great. There was a house who tried really hard to include their very shy new house mate into the group activities but the person never ended up feeling comfortable to join, and it did something to their house dynamic, made it awkward, and they thought it was a shame. There was a house where they were very angry some­one had been placed there, who would be arriving next week. There was even a house who had just had a new house mate placed and they told their new house mate they would all go to the research appointment with us, and then went to dinner with everyone except the new person instead, and left the new person almost in tears waiting for us and profusely apologising.

I ended up writing a chapter in our research report, with many quotes from different houses, and the overall picture, that there seems to be a place for everyone, but finding that place is challenging for 18-20 year olds.

Anonymisa­tion had been guaranteed to the research participants (as we were paid by their landlords) so I couldn’t give them the address where the person was being bullied, as much as I felt an urge to help that 18-year old person who felt very vulnerable to me in that moment that I met them. I advised the person to report their house dynamics, and told them about HappyStudent policy that was in place for this. I also asked the strategists in my next meeting with them to make sure that policy was being upheld. But it left me pondering whether that back-up rule was the ideal solution; it didn’t seem to make anyone happy, only solved a need for the person who got unlucky in the interviews. But is the need for housing, or for housing where you can feel comfortable? Does receiving a room in a house where you don’t feel welcome (or even bullied) do harm or good?


In the end, the research showed that different student houses had different perspectives, and that insiders in the system understood why it was the way it was, and outsiders did not. Housing is an emotional topic, because in the right house you can feel at home, and develop and grow. The right kind of housing is not only affordable but also comfortable in every sense of that word: not leaky, not too cold, not too bare, not too loud, not too quiet, etc. This includes behavior from house mates: if they make you feel uncomfortable, your housing is uncomfortable, even if it’s not leaky and you can afford it.

So HappyStudent kept the policy the same: if you do 20 housing interviews in three months and you don’t succeed, you get offered a place. Our report was used as the foundation for why the policy was reaffirmed. They adjusted some of their communications on their website, and set out to explain better why the policies were as they were, and cautiously added some ideas on what you can do if you feel rejected (try to not take it personally). The report was sent to the city council too. I do not know if they read it, but my impression at the time was that they were satisfied that it had been investigated, and the outcome or follow-up was less relevant to them than the initial outcry had been.

To this day, I still think about this project from time to time, as many other projects that I did. Was it the right research method? Was the time constraint a problem and if so what was it? Did our report calm the tension the organisation was internally feeling about the topic? Did our recommendations help young people feel less rejected after viewings? What could I have offered as policy advice instead? I still wonder if I could have, or one day can, come up with a fresh new idea to overhaul the system and make a better one.

[1] Although the word has the latin root of ‘hospice’ in it, the meaning of the verb ‘hospiteren’ in Dutch is most closely translated to ‘interviewing for a room in a group setting’, not any connotations of care giving, etc. Hospiteeravond = ‘interviewing for a room in a group setting’-evening.

[2] For the record, although this piece is not about the Dutch housing market as a whole, the national housing crisis is currently the biggest it’s ever been; the government estimates 900.000 houses need to be built by 2030 to house everyone. As you can imagine prices are soaring, wait-time is sky-high, etc. The students in this story are relatively lucky compared to many other groups in society, even if their position is not ideal either.

[3] As a private consultancy, GDPR article 9 prevents you from requiring information from ‘special categories of personal data’ for someone to take part in a research, it must be given voluntarily.

[4]  Muslims are missing in this sum-up, for this reason: The Muslim student population in this city often prefers to remain living with their parents until after university, so Muslim students were underrepresented in the shared houses, as compared to in the general student population. We met only one Muslim student and they were not outright biased against others, but they were apprehensive about being accepted as a religious person in what they saw as a secular environment.

Scroll to Top