Human Practices

‘Before you start finding the solution to a problem, you ought to know what exactly the problem is.’ - Dr Prem Jagyasi


Introduction

On our home and project description page we have discussed how the highly pathogenic avian influenza outbreaks in the poultry industry are a very big and complex problem, affecting a multitude of stakeholders. While at first glance you might think that all stakeholders are on the same page when it comes to solving such a problem, it often is the case that certain stakeholders have different needs, interests, and concerns; as with most complicated problems. That is why our team set out to deeply understand the bird flu crisis that the world is now presented with, in order to create a fitting solution which satisfies everyone involved. This is not easy, considering the sheer number of people and organizations involved not only on the local level (The Netherlands), but also on international stages (e.g. EU).

It is also important to involve all the relevant stakeholders in the design process of the solution that you try to create. Even though you understand the problem and the needs of your stakeholders, whether or not they are prepared to use it will most likely still be up to them. Better yet, there is a big chance that you have to consider a whole ‘new’ list of stakeholders because of the nature of your solution. Though your solution may work perfectly in the lab, on paper, or even in practice, engaging with these stakeholders is the only way to truly find out whether your ideas will be accepted and will work in the ‘real-world’.

For these reasons our iGEM team has been very busy with deeply studying and finding out more about the entirety of the highly pathogenic avian influenza problem; from where the problems originate, to who they affect on every level. We incorporated all different aspects of the problem, ranging from economical considerations, to ethical and safety considerations. Only once we have a deep understanding about all aspects of the problem will we be able to come up with a suitable solution. The quote on the top of the introduction by well renowned strategist, writer, and public speaker Dr. Prem Jagyasi beautifully depicts our philosophy when it comes to our human practices approach.

An initial literature review formed the foundation of our research, which we then used to identify crucial and relevant stakeholders, as well as many experts in the field. We aimed to reach all these stakeholders to incorporate their views and values to the overview we have built up of the problem. We used these large quantities of high-quality interview-data to find and create a safe, responsible, and sustainable solution that is most fitting to the problem. These stakeholder interviews also nudged our activities and research into the right direction. A crucial stakeholder in our solution is the general public, which should not be overlooked. Therefore, as a result of our research, we decided to include a survey on avian influenza in the poultry industry and GMOs, in the light of our solution. Moreover, we published workflows and templates to effectively and responsibly deploy surveys to efficiently gather data for future teams to use in their human practices work. All of our gathered data was not just used to create a safe and sustainable solution, but was also employed to gain insights into the future perspectives and responsible implementation of our strategy. The information provided to us through our research helped us incorporate several Safe-by-Design principles into our project, further ensuring the safety of our proposed solution.

Through our human practices, we show our methods and what we undertook to make Nanobuddy as practical, responsible, and safe as possible. Our integrated human practices show how we applied and integrated our gathered knowledge, to ultimately utilize it to truly make Nanobuddy a project which has a positive impact on the world, and everything living in it.

The problem in a nutshell

The Netherlands is plagued by the spread of the highly pathogenic avian influenza (H5N1) with record breaking numbers of infections in both the industry, as well as wild bird populations. Avian influenza is becoming an increasingly large and dangerous issue in not just The Netherlands, but in the whole world, and should be dealt with accordingly. Unfortunately, current regulations make it very difficult to implement strategies such as vaccination to tackle this epidemic. These are a result of EU regulation which is meant to stimulate trade; vaccinated meat is difficult to distinguish from infected meat. Not vaccinating is simply just way more easy. As a result of this trade-stimulating policy, farmers and local governments are left with only one option: biocontainment. As it drastically decreases costs on vaccination, and stimulates trade, this might sound like a good solution. However, considering the rapidly increasing avian influenza cases it appears that the strategy of keeping the virus out is not sufficiently mitigating the problem. Biocontainment strategies basically come down to minimizing the chance of your poultry flocks catching the virus. This is achieved through rigorous biosecurity protocols, but also through the (preventively) culling of poultry, caging of poultry, and transport restrictions for the industry. When your biocontainment strategies are not working efficiently enough to keep the virus out, is culling all of your birds still really the best way to combat spread of this disease? On top of that, immunologists are sounding the alarm bells that the virus might jump to humans if left uncontrolled for too long. Without a definitive solution on hand, the poultry industry, NGOs, and political parties alike are pleading for the introduction of vaccination strategies. Vaccinations however, are also suboptimal as the only suitable vaccinations for this highly pathogenic variant have to be introduced intravenously; through an injection, often more than one. This is different to vaccination strategies for lower pathogenic diseases such as Newcastle's disease for example, where just a few spray vaccinations is sufficient to proficiently protect poultry. Considering we would have to conventionally vaccinate one hundred million poultry in The Netherlands alone, there has to be a better way to swiftly protect our poultry sufficiently.

In the project description page, we have included an elaborate problem description which we created after doing an extensive literature review. Here we will briefly address the problem that we are trying to solve with our solution. For references and more information we refer you to the problem description on our project description page.

Our Approach

To gain an in-depth understanding of the problem of avian influenza outbreaks in the poultry industry, we consulted both literature and involved stakeholders. It is very important to us that our solution actually touches upon the wishes of our potential end-users, and addresses the problem appropriately. Therefore, it is only fitting to make stakeholders one of our largest sources for information regarding the avian influenza crisis, and for feedback on our solution. The literature review that we conducted can be found on our project description page.

Stakeholders often have knowledge of the latest updates, current trends, opinions, needs, and other not so obvious stakeholders that are working on, or are affected by the problem. This is highly-valuable information that a literature review simply cannot provide you. To ensure that we included all stakeholders, we had thorough brainstorm sessions and conducted an accompanying stakeholder analysis. After identifying every stakeholder that we would need to contact, we set out to organize an interview with every one of them to investigate their perspectives and gain new insights. To obtain the opinion of one of our potentially most influential stakeholders, the general public, we developed a survey later on in our human practices work.

Furthermore, we gained feedback on our project through the numerous collaborations we did with multiple teams. For example, we attended multiple iGEM meet-ups where we presented our project, answered questions, and received points of attention from other iGEM teams. More on this can be found on the integrated human practices page, the partnership page, and the collaborations page.

Altogether, this page describes the steps we took to identify, approach, and engage with stakeholders in a responsible manner. Moreover, we show how to efficiently collect data gathered/obtained from stakeholders. Information on how we put the data we gathered in our human practices to use can be found on our integrated human practices page.

Stakeholder analysis

As explained in the previous sections, it is very important to include as many stakeholders in your research as possible. More stakeholders, means more perspectives, and valuable feedback. We organized rigorous brainstorming sessions to identify as many, potentially contactable, stakeholders as possible. To guide this process, we conducted a stakeholder analysis.

We wanted to identify every stakeholder that is involved in the problem. Therefore, we first wrote down as many actors as possible on various post-its and excel sheets. When we were done, we ended up with a comprehensive list of the stakeholders that are involved in avian influenza outbreaks in the poultry industry: a good place to start when preparing for our human practices work.

Brainstorming

We first started by organizing a brainstorming session with the people involved in our human practices work, where we would write down every stakeholder we could think of through post-its. After identifying the more obvious stakeholders involved in the problem, we did some research on the internet to identify those that were less obvious. The final collection of stakeholders was then put together in different groups (e.g. government, industry, and other different societal groups). We then created an overview of all the stakeholders we identified, and the larger group they belong to. This overview can be found in figure 1, and formed the first stage of our stakeholder analysis. What we immediately noticed is that a large proportion of stakeholders fall within the ‘industry’ and ‘government’ groups.

Figure 1. The collection of stakeholders involved in the avian influenza crisis, and the group they belong to, in one overview: the first stage of our stakeholder analysis. Made using Biorender.com.


Analyzing

The next step was to identify the most relevant and important stakeholders. These would eventually be the stakeholders that needed to bo spoken to as soon as possible. Also, it is important to have a good understanding of how different stakeholders relate to each other. For this, we used a popular stakeholder analysis tool: the stakeholder matrix. This tool allowed us to separate our stakeholders into different groups, which shows us how we would have to interact with them.

You can find different variations of the stakeholder matrix, but eventually they all come down to the same main principles. One axis representing the interest, stakes, or availability; in our case how invested the stakeholder is in the problem. The other axis, representing the power or influence; for us this shows the ability of the stakeholder to facilitate change to act on the problem. Every stakeholder gets appointed an x and y value, and therefore a place on the stakeholder matrix spectrum. Eventually we were met with the stakeholder matrix as shown in figure 2. The matrix is divided into four separate grids, in which the stakeholders from different societal groups reside. These grids show how you need to take into account each stakeholder in order to be most successful in creating an effective and useful solution. Across all different stakeholder analysis techniques, these four grids usually come down to this:


Figure 2. Filled in stakeholder matrix for all stakeholders involved in the avian influenza crisis. The x-axis represents the interest/stake in the problem, and the y-axis represents the influence/power of the stakeholder. Made using Biorender.com.


The final stakeholder matrix we landed on constituted of all of the stakeholders identified in figure 2. It was interesting to see that most of the stakeholders with a high interest are almost all either the animals themselves, or the industry directly involved with poultry. In the ‘engage closely’ grid we have identified NGOs, larger industry organizations, and mainly acting governmental institutions to be our most important stakeholders. One notable stakeholder is the Wageningen Bioveterinary Research facility which is in close contact with the government with regards to avian influenza. Since they are a well respected research institution within the field of agriculture, and directly feed into the relevant governmental institutions, we decided that they merit a place in the top right of the grid.

We thought it was also noteworthy that both the pharmaceutical industry, investors, and a lot of governmental departments appear to have a low interest in avian influenza. This is be because of that it is not a human epidemiological threat right now, and that vaccinations in the animal industry are not yet allowed because of regulation. Currently, these stakeholders do not appear to have a large interest in addressing this problem. This leads to an interesting dynamic where both parties have less interest in the problem because of each other. Legislatory stakeholders are less likely to change policy for a solution that does not yet exist; and actors like pharmaceutical companies are less likely to develop a drug which is not allowed to be used. For these reasons, we fit them into the ‘keep satisfied’ grid.

We also ended up creating a stakeholder map as a result of the interviews we conducted. We realized that the poultry industry alone is a very complex system with a lot of moving parts. To keep an overview of how all of our stakeholders interacted with each other, we decided to create a stakeholder map which we regularly updated throughout our human practices work (figure 3).


Figure 3. Stakeholder interaction map showing an overview of how stakeholders in the avian influenza crisis interact with each other. Created using Biorender.com.


Overall, the brainstorming and stakeholder analysis tools helped us identify all of our relevant stakeholders, and helped us categorize them with respect to influence and interest in the problem that we are trying to solve. The analysis helped us with weighing all the different needs and concerns of stakeholders with regard to the solution we came up with. Throughout our human practices, we learned more about how these stakeholders interact with each other, which is represented in figure 3.

Figure 3 shows the intricate relationships between various stakeholders. Mainly, we focussed on the poultry industry itself. As a whole, the poultry industry was different to what we imagined it would be beforehand. You can see that we have build up the industry as a pyramid, which refers to the amount of poultry found in each sub-sector. The few poultry at the top of the pyramid represent the grandparent stock which form the basis of the whole industry. These poultry are then sold to breeding companies who breed them, and sell their eggs to hatchery companies. As soon as the chickens are hatched and are a bit older, they will be sent to the broiler farms that people are usually more familiar with. We were actually surprised to learn that the poultry industry generally consists of this pyramid structure, and we carefully considered this structure when designing our implementation for example.

Besides the poultry industry (and retail), we have also decided to add the three other main players in this problem: government, researchers, and the general public. To briefly summarize our design choices in figure 3, we included only these specific main players as that would give us the most concise and structured overview of the relevant people and organizations in one figure. A figure containing every individual stakeholder would simply be too overwhelming, and would make our understanding even more vague. As a whole, the general public is mainly involved through their function as a consumer, and therefore only interact with the other stakeholders through the retail industry. There are some exceptions such as hobbyist poultry keepers and restaurant owners, who interact with other stakeholders because of their occupancies as shown in the figure. NGOs and legislatory governmental bodies mainly interact with the poultry industry indirectly, trough union organizations or pharmaceutical companies for example. Acting governmental bodies more often directly interact with the poultry industry through bodies like the NVWA, who are responsible for diseased animal disposal and avian influenza biosecurity. Researchers, experts, and veterinarians in general interact with all of the stakeholders. Providing information and knowledge on which the stakeholders are able to act.

Together with the already carried out stakeholder analysis, we now had a well rounded understanding of how we should incorporate the feedback and weigh the needs of stakeholders we discovered throughout our human practices.

Stakeholder engagement

With our stakeholder list ready to go, it was time to get out of our office/lab and consult our stakeholders. By now we have already compiled a pretty good idea of what our solution was going to look like by brainstorming with the team and discussing it with our supervisors. Our plans meant we had something to show to our stakeholders in order to get proper feedback. In the following paragraphs we will explain how we responsibly contacted our stakeholders, and the methods we used to gather valuable data which would ultimately shape our project to be safe and implementable in the ‘real-world’. The latter of which can be found on our integrated human practices page.

Stakeholder approach: from initial preparations, to the interview

Preparing for your stakeholder interviews can take a lot of time, as it is crucial that you take into account all the relevant ethical guidelines, privacy laws, and responsibilities as a researcher conducting these types of research. Luckily, there have been a lot of iGEM teams which have put a lot of time and effort into their stakeholder approach. Our predecessors in particular, iGEM Groningen 2021, provided numerous templates, workflows, and guidelines based on their extremely intricate stakeholder approach. Where the 2021 team had to sacrifice quantity over quality for their stakeholder interviews because of their efforts to reach out responsibly, we were able to use their contributions and consequently provide both quality AND quantity; taking our human practices to the next level. We even managed to expand on their work, creating more useful human practices workflows and templates for future teams to use. If you are a future iGEM team, we highly recommend you also check the Groningen 2021 page when designing your own human practices approach. In the following paragraphs we will discuss how we designed our stakeholder approach.

What do we want to know?

You have to figure out what you want to know from the stakeholders exactly before you actually go to speak to them. From previous experience with interviews, our team quickly decided that we would design a standard question list that we would use for every individual stakeholder. This way, it is also very easy to compare insights, needs, and other data gathered between different interviews. Still, the interview would be conducted through a semi-structured manner which encourages two way communication; very important for these types of interviews. We compiled a question list which we would later send out to interviewees beforehand. Of course, sometimes we added a few new questions because of the nature of the interview. Sometimes it also happened that we quickly ran over questions which were not that relevant to specific stakeholders. However, still asking them is important since it might uncover insights which you were not previously expecting! The questions we compiled were aimed to gain an understanding of the problem itself, already existing solutions, and our own solution.

For our stakeholder approach strategy, we also decided to ask about other stakeholders during the interviews. These questions were aimed to further deepen our understanding of inter-stakeholder relationships, but also to potentially discover new stakeholders. Through our so-called ‘stakeholders-know-more-stakeholders’ approach, we have identified and spoken to more stakeholders than we originally anticipated, showing that this strategy could be interesting to use for iGEM teams in the future! Along with the questions, we added the structure of the interview (almost like an agenda) so the stakeholders knew perfectly what to expect beforehand. The interview overview we sent to stakeholders can be found below in figure 4.



Last preparations: reaching out responsibly

Before we could start reaching out to our stakeholders, we would need to ensure that we would do this responsibly. When interviewing people you have to consider that you are doing research using actual living human beings that have a lot of rights, and more importantly, feelings. To ensure that we respected the relevant laws and regulations in place to for example protect the privacy, rights, and data of research participants, we did research on how iGEM teams in the past did this. Rather quickly, we found that the best way to ensure that your human practices work is responsibly conducted, is by focussing on two main concepts: data protection & informed consent.

We started out by making important choices about what data we were going to collect, from whom we were going to collect it, and how we were going to approach our informed consent form. We did this through the framework described by iGEM Groningen 2021. A graphical summary overview of the steps we took to ensure data protection and informed consent for our own research is attached in figure 5 below.


Figure 5. The graphical summary of the steps taken to ensure data protection and informed consent for the stakeholder interviews. Inspired by the data protection and informed consent steps and considerations described on the iGEM Groningen 2021 human practices wiki. Made using Biorender.com.


We first started out by consulting the relevant legislation with regards to data protection and informed consent. We started, just like many other iGEM teams, at the European General Data Protection Regulation (GDPR), which has been introduced as of May 25th, 2018 to harmonize data privacy laws across Europe [3]. The GDPR states how to ensure data protection, as well as informed consent for research such as stakeholder interviews. In addition, we read up on the possibility to appoint a data protection officer (DPO) as defined in the GDPR [3], or whether we could conduct a data protection-impact assessment (DPIA) [4]. Keeping all this in mind, we started making the important choices regarding data storage and data collection.

To start off with, we looked at what type of data we would like to collect in order to improve our solution. Since we are working on avian influenza, we did not have any reason to interview stakeholders who were at risk of not being able to give proper consent. These vulnerable research participants could for example be minors and patients in the medical system [5]. We decided to not interview this group to make the process of creating our informed consent sheet smoother, as well as minimizing any additional ethical concerns we would need to take into account. These risk groups did not come up in our stakeholder analysis. As such, we did not see a reason to include them. For the survey we conducted later on in our human practices, we decided to do the same. In our educational outreach we did interact with and taught minors in synthetic biology. We only did this after careful consideration, consent from the school we attended, and consent from a parental figure.

On top of that, we decided to never collect any personally identifiable information in our notes, and we did not write about potentially sensitive information. We did not see a reason why we would need to include the names of our stakeholders and experts. We decided to refer to our stakeholders using gender neutral pronouns (they/them), unless they preferred otherwise. We called the stakeholders by their expertise, occupancy or affiliation, and position within the industry/company. Before finalizing the summary we had written of every interview (including the expertise/position we would use to refer to them), we sent it to the stakeholders for confirmation so that we were allowed to publish it, as well as a chance to change statements we wrote down.

During data collection and storage, you often use a lot of digital technology to make appointments, set/conduct meetings, and reach out. We ensured that we always used trusted password protected platforms with multifactorial authentication. On top of that, we deleted everything that we would not need in the future. This includes, but is not limited to, notes, recordings, and emails containing informed consent forms. We also went the extra mile by deleting meetings from our calendar, deleting phone calls (if contact was made through the phone), and every email we did not need in the future.

Since it is important that a participant is aware and well-informed about the research they are participating in, we also decided to include an information sheet together with our informed consent sheet (figure 6). This information sheet contains all the relevant information that the participant should know. These are the main subjects touched upon in a proper information/informed consent sheet:



After summarizing all of our plans so far, we decided to create our information and informed consent sheet. We used the template provided by iGEM Groningen 2021 and went to work filling it in to make it fit our own research purposes and goals. We did a final review and checked whether we were still complying with the GDPR. After that we also found that there was no need for a DPIA, nor the appointment of a DPO. We contacted the data protection officer and ethical committee from the faculty of social and behavioral sciences of the University of Groningen for some last feedback. After the last round of feedback from our supervisors we had finalized our informed consent and consolidated our data protection strategy. The final information and informed consent sheet which we used for our research can be found below in figure 6. In this sheet, all the above information is addressed and extensively explained/ elaborated on.



Step-wise plan of approach: contacting stakeholders

With our informed consent and data protection strategy well defined and our questions ready, we decided to reach out to every stakeholder we have identified as soon as possible, prioritizing our interviews according to the stakeholder matrix (figure 2) from our previous analysis. We followed a step by step approach, which was largely adapted from the Groningen 2021 team with a few personal tweaks (figure 7).


Figure 7. Step by step process describing how we approached stakeholders, up until everything we did after the interview (including our very own ‘stakeholders-know-more-stakeholders’ approach). Made using Biorender.com.


The first step to initiate contact almost exclusively happened through an email where we would introduce ourselves, and the research that we were doing. We would explain why we would like to talk to this specific stakeholder, and what they could contribute to our project. We aimed to keep the emails short and concise, so not to scare away potential interviewees with a very long email. If the stakeholder wanted to learn more about our project, we added a short abstract of our project to the same email. We also made it our goal, even though we used a script, to make the emails as personal as possible in order to attract more responses. The script that we used, and the abstract we sent with the emails can be found in figure 8 and figure 9 respectively.




Preparing for the interview consisted of picking a date, place to meet (in real life or online), and sending our informed consent/ information sheet which the stakeholder needed to sign before we conducted the actual interview. It is very important that this happens beforehand, as the stakeholder should be well aware of their rights, and what to expect during, and after the interview. We also send the stakeholder the questions/topics that we would like to discuss (figure 4), so that they know what to expect and can prepare accordingly.

Before the interview, we join the online meeting or open the reserved room well in advance. In figure 10 we have added a picture of what most of our meetings looked like. During the interview, we use the question file (figure 4) to guide the interview. We make notes and a recording for which we first, again to be sure, ask verbal consent. We carefully go over the important parts of informed consent and data protection to make sure the stakeholder understands all the following steps after the interview, and is still okay with everything. During, but also after the interview, we asked questions about other stakeholders which we might have overlooked before. Often our stakeholders came up with people that we had previously not identified in our stakeholder analysis and they were able to refer us to them. This ‘stakeholders-know-more-stakeholders’ approach provided us with the opportunity to speak to more stakeholders in a circular manner as visually represented in figure 7.


Figure 10. Ronald, Yasmin, and Mink getting ready for an interview with a stakeholder. Note that we had to make it anonymous because we do not share personally identifiable information.


After the interview, we wrote the summary based on the AREA framework of iGEM Exeter 2019. We will explain these summaries further in the next section (stakeholder interviews: gathering and processing new insights). We asked the stakeholder to confirm this summary, after which we uploaded it as a password protected pdf file on a password protected server with multifactorial authentication. This was also the moment where we asked stakeholders whether or not they would like to remain involved in the research project. This meant that we would send them updates on how everything was progressing every once in a while through a brief ‘newsletter’ email.

Stakeholder interviews: gathering and processing new insights

We organized our interviews so that there were always at least two people present: one interviewer, and one dedicated note-taker. This way we made sure that the conversation could continue smoothly. To ensure that we did not miss anything in our notes, we also made recordings which could be used internally alongside the notes we took. The interviews were designed to take between 30 and 60 minutes, but in reality they often took about 90 minutes (if the stakeholder had time). You can imagine that you talk about a lot of topics and questions in all this time, especially in semi-structured interviews. To make sure that we use the data we gather in our human practices to improve on our project in a structured and organized manner, we used the aforementioned AREA Framework for Responsible Innovation, created by Professor Richard Owen, as introduced by iGEM Exeter 2019 (figure 11). This framework is designed to facilitate opportunities and creativity for science and innovation [6]. The AREA framework is often used by research councils to encourage a responsible approach to innovation. We have attached a graphical overview of the AREA framework as defined by iGEM Exeter 2019



Figure 11. The 4 steps of the Anticipate, Reflect, Engage, Act (AREA) framework including the extra layer iGEM Exeter 2019 introduced: description, contribution, adjustments, and our next steps).


The 4 steps in the AREA framework especially served us well in structuring our integrated human practices. Just like iGEM Groningen 2021, we decided to fill in our AREA frameworks more elaborately, sticking to our high quality and high quantity approach. Since we decided not to include any personally identifiable information into our AREAs, we were unfortunately not able to include pictures of the meetings we had with stakeholders. This tool not only helped us structure and organize our thought processes, but also facilitated new idea’s which we could then put to use in our project. To see how we used this framework to adjust and improve our project, look on our integrated human practices page!

The survey

Our most influential stakeholder yet?

Quickly after conducting our first interviews, we found that besides the political hoops one has to go through when it comes to protecting poultry against the virus, there is also a large influence of public opinion. Multiple stakeholders addressed the matter of public opinion to us, which we duly noted in order to potentially consult this ‘stakeholder’ in the future. At the beginning of our human practices work we decided to sideline the ‘general public’, and focused more on speaking to as many experts as possible. However, considering the sheer number of stakeholders which addressed this issue to us so early on in our interviews, we decided to look into how we could gain the public’s perspective.

The matter of public opinion is especially relevant when it comes to the usage of GMOs in the food industry. Opinions on these topics are often polarized, and we were very eager to find out what the public opinion would be on our project. The influence of the public opinion is large, and considering every person that eats poultry on a regular basis is considered a stakeholder, could mean we have found our largest stakeholder yet. For this reason, we wanted to test the waters and find out what the public opinion is towards usage of GMOs to prevent spreading of the avian influenza virus among poultry.

For the complete version of our survey, and the results, check out our survey page. Check out our integrated human practices page to see how we used this data to shape our project even further!

Developing a responsible and valuable survey

When using a survey in any project, it is important that they are implemented with great care. Just like any other experiment you would do in the lab, you have to be very careful and thoughtful with its design and while gathering data. Poor execution of an experiment might lead to poor or even useless results; poor execution of a survey will almost always lead to results that do not accurately reflect the opinion of the participant. This would be irresponsible as we are working with real people, and asking them for their time by participating in our research. Since we are working with human participants, we also have to be very careful with regards to data protection and informed consent. Therefore, we took great care in designing our survey to be responsible, while making sure it will provide us with valuable information.

Here we will describe the 10-step process we took to make our survey responsible, and explain the precautions we took to ensure the survey would provide us with valuable data.

Just like any other contribution to a research project from a participant, you will absolutely need to ask them for informed consent before participation. When responsible for conducting a survey, this is one of the first things to start with. Thanks to the amazing human practices framework put together by the iGEM 2021 Groningen team, we have already created an information and informed consent sheet in our earlier human practices work which deals with all the important ethical, legislatory, and institutional considerations relevant to stakeholder interviews. For a more in depth explanation on how to write a proper informed consent form, look at our stakeholder approach. Here we will briefly address the relevant changes we made to the already existing document.

Since the survey is a different type of data collection, we created a new shorter informed consent/information statement which is fitting for survey-data collection. Since we will not collect any potentially personally identifiable data, shortening the statement a lot easier. We do have to ask participants for an email address which of course never will be published. We only need it so that we are always able to remove a participant's data if they request it. Since this informed consent is largely based on EU legislation and assumes an age of 18 and older, we also added a statement that the participant was an EU citizen and 18 years or older which the participant had to confirm. We also clearly stated how, and where the participants' data was collected, and the precautions we took to protect their data/privacy in a shortened version of our information sheet. This is especially important because we do have to ask participants for their email address which is sensitive data. We ended up saving the data through a password and multifactorial authentication protected server through a trusted and reputable provider of survey tools, which could all be found back in our informed consent/information statement.

Since we based this informed consent/information statement on the informed consent and information form we already had, we were confident that we included everything. We took the same steps as we did when we created our main information sheet/ informed consent sheet. Still, we double checked with the regulations within the university and the GDPR act from the EU [3]. We also checked in again with our supervisors for additional feedback The shortened, more specific informed consent/information form that we eventually created can be found in figure 12. In this figure, the above steps are extensively and elaborately explained.

Figure 12. Informed consent/information statement that we placed at the beginning of our survey. Important storage places are blocked out because of safety reasons.

We developed our survey in order to gain an insight on the public opinion towards avian influenza, vaccination, and GMOs, in the light of our solution. Therefore, the answers to the questions that we will present to participants will have to accurately reflect their opinion on these matters. The aim of our survey was to find out the following:

  • How concerned is the Dutch population about avian influenza?;
  • Is the Dutch population concerned about vaccinations in the animal industry in general, and does this change for avian influenza specifically?;
  • Would the Dutch population be accepting of our GMO viral protection strategy in the animal industry in general, and would this change for avian influenza specifically?;

Through asking about vaccinations and avian influenza in general, we hoped to find out whether a dislike or approval of our project might be a result of a general dislike towards vaccinations, or for example a general interest in avian influenza. In addition, we could find out what they thought of our solution in contrast to conventional strategies.

We hypothesized that the general public would not be too fond of our alternative protection strategy as it involves GMOs in the food industry, something that often is not looked kindly on in the EU, as well as The Netherlands. However, we also suspected that this might change for the better if the participant worries more about avian influenza outbreaks, and would therefore consider our strategy earlier.

For us the target group is relatively straight forward: the general public. This presents us with a few complications though:

  • The survey has to be relatively short (5-10 min), as we need a lot of responses to obtain valuable insights. A short survey is crucial when wanting to obtain information from complete strangers.
  • The survey has to be understandable to the general public. In the design of our survey, we have to consider that everything needs to be clear to every possible reader. If a reader cannot properly understand a question, they cannot properly convey their opinion. Points of attention were not to use jargon, and we decided to include two neutral text boxes that explained a few key concepts (e.g. how our solution works, without explaining why we chose for this approach or why it might improve upon already existing prevention strategies).
  • If we want to accurately reflect the Dutch public opinion, besides other countries, we would have to translate the survey to Dutch so as not to exclusively include English speaking participants. Having the same survey in multiple languages can lead to data biases since certain phrases or questions might be interpreted differently in different languages (even though the translation is still spot on). Since this is very difficult to control for, and we want to accurately compare responses between participants, we decided to only provide the survey in Dutch.

To determine the sample size that we should aim for, we used a commonly used formula in statistics [1]:

Figure 13. Formula to calculate sample size for a research project. p = population size, e = margin of error (percentage in decimal form), z = z-score.

We used a z-score table [1] to determine our z-score (easily found through an internet search) for a desired confidence interval of 95%. In our case, z = 1.96. For our population size, we used the latest data available on the Dutch population size: p = 17.749.262 [2]. Finally, our margin of error was set to e = 5.

To conclude, according to our calculations we should aim for a sample size of 385 participants.

It is important that you choose your delivery vehicle with the purpose of your survey in mind. An extreme example, you should not spread a survey about the public opinion towards the meat-industry at different vegan food stands. You have to tailor your delivery strategy in order to accurately reflect your population. We had to think of a way we could gather data from the Dutch population which would not be biased and would accurately reflect the general public opinion. We chose an exclusively online survey through a trusted platform. We spread our survey through the following networks and made sure that we did not over-promote the survey to specific demographics (minimizing biases):

  • Flyers with QR codes spread around in public places with diverse demographics
  • Online survey sharing platforms
  • Social media
  • Promotion through our communication/educational outreach (e.g. the blog)

This is easier said than done, as one of the key values of our survey was that it should be relatively short (5-10) minutes. For questions regarding an opinion, we decided on an agree-disagree range consisting of 11 values (0 representing ‘completely disagree’, and 10 representing ‘completely agree'). For our applications this was the easiest way to accurately and efficiently study and compare different answers by means of visual aids and potentially statistical tests.

We also decided that we would ask participants for the following non-personally identifiable data:

  • Age
  • Gender
  • Living area (which specific Dutch state)
  • Educational level

It is important to consider what factors might be able to heavily influence your data. Therefore you should almost always ask questions about the participant’s gender, age, living area and educational level. You therefore have to make sure that your sample contains participants that are well-divided across the different demographic groups. This last question is especially important since this is likely to heavily influence the opinion of a participant in our other questions; a question that we only put in after carefully considering the contents of our survey.

We then decided to also include another statement in the informed consent sheet. This would say that: ‘there are no right or wrong answers, please enter the first thing that comes to mind after understanding the question’, which was aimed to promote participants not to dwell on questions for too long, and provide them comfort when giving us their opinion.

To test whether we could obtain results that were able to answer our research questions, we conducted a hypothetical responses test. Using this test we gathered data from 500 imaginary participants with randomized answers. We then got to analyzing our results, which kind of felt like doing a general repetition for a play. Rather quickly, we found out that we actually missed key questions in order to properly answer research question. This hypothetical experiment not only helped us gain a better insight in what types of statements we could make using our results, but also helped us prepare for the real analysis of the actual survey respondents’ results. This was also the moment where we realized that some questions, were directly targeted at people that eat meat. We therefore added another question regarding our stakeholder demographic: amount of meat consumed on a weekly basis.

After carefully creating a final list of questions that gave us an insight in the participant’s opinion on the presented topics, we had to ensure that the questions are understandable, accurately address our research goals, and cannot be interpreted multiple ways by the general public. This would go a long way to improving our survey, and will help us gain even more valuable data to use in our project. On top of that, we had to ensure that our informed consent statement was understandable since it is crucial that a participant knows what they are in for and what their rights are. Basically, getting as much feedback on your survey set-up is crucial to ensure it’s responsibility and accuracy. We used two methods to ensure that we executed the survey to the best of our ability so that it has maximum value:

Peer review

The first method we deployed was a peer review by team members not involved in the development of the survey, and our supervisors. This review was aimed at grammatical errors and spelling mistakes, whether we address our research questions or not, and overall structure. After a few rounds of gathering feedback, and incorporating it in a new draft, we moved on to the next stage.

Focus Group

In order to make sure our survey draft is understandable, and to get an insight on how people interpret our questions, we conducted a focus group. It is important that this focus group accurately reflects the people in your population. No overrepresentation of certain demographics was therefore considered when selecting people for our interviews. We managed to get a focus group together consisting of a small sample of the Dutch general public. We talked to 3 females and 2 males (n = 5). Mean age was 49.0 years old (SD = 8.05). All participants finished their post secondary education (Dutch MBO / HBO). Overall we felt that we have received enough feedback to continue, and since the focus group is diverse, we felt that this focus group was sufficiently representing our sample. Below we will discuss how we structured our focus group, what insights we gained, and how we adjusted our survey accordingly.

Methodology

Before participants of the focus group got interviewed, we asked them to read the information sheet and informed consent form we specifically created for the focus group (which was based (almost 1:1) on our stakeholder interview informed consent form found in Figure 6), which we needed filled in before gathering any data. All interviews with participants of the focus group were conducted separately, so that their answers and thought processes would not be influenced by others. The researcher had an observing role as participants worked through the survey. This means limited ‘small talk’ and ‘explanations’ which might influence how they would fill in their survey in private. An important aspect of our focus group is whether the informed consent statement in our survey draft is written down clearly as well. First, we let the participants read the draft survey informed consent statement very carefully, and asked the participants a list of questions to confirm that they properly understood their rights, and all the other information as stated in the GDPR act we need to provide was understood too. Then, we asked participants to fill in the draft survey while thinking out loud, so the researcher could follow the thought process of participants. This way we can ensure that our questions were properly interpreted, meaning the answers of participants actually represent the opinion that we would anticipate with a given answer. We also asked for general feedback on the draft survey from the participant perspective.

Results and feedback

According to all stakeholders, our informed consent statement at the beginning of the survey was understandable. They also confirmed that they were properly able to perceive all the relevant information that this piece of text offered. Therefore, after confirming all the elements named in the GDPR were understood, we concluded that participants of the survey are properly informed after reading the informed consent statement, and checking off that they agree to participate in our survey. Then, we proceeded to the actual survey.

Right off the bat, we realized that we would have to adjust and review our survey even more thoroughly. Without going into specifics, we for example quickly realized that one of the first questions, such as ‘diet: 1) no restrictions, 2) vegan, 3) vegetarian, 4) other …’, can have different interpretations which we definitely did not anticipate. This question was put in to understand what the stance of a participant is towards eating meat, which likely heavily influences results on questions about the meat industry later on. One participant would have answered this question with ‘4) other …’, reasoning that they answered this because they ‘had a very specific diet, which was not represented’. However, when asking for more detail, this diet had nothing to do with their meat intake, or the type of meat they ate. Another participant stated that if they would have had a gluten allergy for example, they would also have put down ‘4) other …’. This question alone clearly underscores the importance of conducting this focus group, as we now knew to specify the ‘diet’ question to be something like ‘diet regarding meat intake’.

Another notable example of misinterpretation was a question about their stance on meat coming from GMO treated livestock. The way this question was originally formulated, allowed for lots of differences in interpretation. We designed the questions in a way that participants could answer on a scale of 1 (strongly agree) to 10 (strongly disagree). For this question, and many others, people that shared the same opinion ended up giving answers which are polar opposites from each other (e.g. 2 and 8) due to various reasons. Similarly, people that for example: ‘were pro GMO’s’; ‘ were anti GMOss’; or ‘did not really care’, all answered similarly for the same question (e.g. 3, 2, and 2). This was definitely not anticipated and means that answers from these questions cannot yield reliable results; unless reformulated.

Participants overall did not feel like they were steered into giving specific answers due to the questioning style, order, or content of the question. However, some participants stated that the two information boxes with background information on our project and the avian flu crisis might influence questions on their personal concerns about the avian flu crisis. Still, other participants stated that these information boxes helped them understand the rest of the survey better. Participants also stated that the order of the question might be altered in order to make the survey flow better.

Adjustments

We adjusted problematic and unreliable questions according to the feedback that we received. After some feedback on the structure and arrangement of the questions, we also decided to rearrange the questions accordingly.

We decided to drastically change the diet question. We have had feedback stating that ‘no restrictions’ and ‘actively eating less meat’ might lead people to lean towards the latter, while they actually do not actively restrict themselves in terms of meat consumption. To make sure we get the best insight into the meat consumption of participants, we will ask them about the amount of meat they eat each day. If none, we ask whether they are ‘vegetarian’ or ‘vegan’. This allows for a more clear gradient between a ‘no restriction’ and ‘less meat’ individual, easing the answer process and reliability of the survey.

It is difficult to make the choice about whether or not to include the explanatory text boxes with additional information. On the one hand it helps people to properly understand our survey, improving the reliability of the provided answers. On the other hand, it might influence people to be more keen to score differently on questions about their personal concern on avian influenza. After careful consideration, and after discussing it with our supervisors, we decided to partially delete the information provided. We made this information neutral and strictly informative about the concept of our strategy.

After the adjustments, we can expect that the results better reflect the opinion of people that we would anticipate.

For us this was the time to review the whole process of creating the survey, keeping the feedback in mind. We made the last changes to finish the final survey, and published it through the aforementioned networks. Please find our finished survey and the results on our survey page!

We regularly checked whether we had overrepresented demographic groups responding to our survey. If this was the case when the survey was still open to responses, we would be able to respond to it. The only difficulty we found was getting enough responses to match our desired sample size, and over/underrepresented Dutch states. We responded to this by actively spreading our survey more frequently, and specifically more frequently in the underrepresented states.

It is very important that you consider all the data you have gathered in your survey, even if they do not show the results that you anticipated beforehand. Accurately describing and showing your findings might be the single most important aspect of conducting a survey. It is also important to include a figure which shows the specifics of the demographics that you reached through your survey, in order to provide evidence that your data might or might not be biased. Showing this data to the reader is important as well. We used various methods to show our (demographic) data visually, and used statistical tests where appropriate. Our data, and how we chose to visualize it, can all be found on our survey page.


Materials for future iGEM teams to use

Often when you are working on a iGEM project, or any long-term project for that matter, you forget the crazy hours you had put into preparation, background research, and optimizing your methods. Even though the work of previous teams allowed us to move swiftly through a lot of the steps involved in working responsibly throughout our human practices, preparations took a long time. When working on our own human practices we also realized that we had done a lot of work behind a desk to prepare or perfect our human practices approach, rather than actually interacting with our stakeholders. We have built upon previous teams contributions by further expanding the tools they provided to us, as well as creating new tools and workflows ourselves. On this page, we would like to address and share these contributions, so iGEM teams in the future can have even more time to step out of the lab/office and interact with their stakeholders.

Informed consent statement fitting for gathering data through surveys

The informed consent that we had created based on the template provided by iGEM Groningen 2021 was exclusively focused on gathering data from stakeholder interviews. We have studied this template, and modified it to be applicable and appropriate for gathering survey-data from research participants (figure 14). We hereby not only contribute something new for future iGEM teams to use, but also build upon already existing work by other teams, making their initial contributions more widely applicable.



10-step workflow for efficiently developing and deploying a responsible and valuable survey

We have learned a lot from previous iGEM teams on how to responsibly gather and store data from stakeholder interviews. We wanted to build further upon the contributions of these teams by creating a workflow for responsibly gathering and storing data from surveys (figure 15). We think that this workflow provides a good new tool for iGEM teams in the future that are working on their human practices. We hope this framework will guide them to easily implement surveys into their project, as they are a very powerful tool to use for any iGEM team.


References

[1] Kadam, P., & Bhalerao, S. (2010). Sample size calculation. International journal of Ayurveda research, 1(1), 55–57. https://doi.org/10.4103/0974-7788.59946

[2] CBS: Centraal Bureau voor de Statistiek. (2022, 25 februari). Bevolkingsteller. Accessed on August 1 2022, from https://www.cbs.nl/nl-nl/visualisaties/dashboard-bevolking/bevolkingsteller#:%7E:text=Eind%20augustus%202022%20telde%20Nederland,maand%20staan%20in%20deze%20tabel.

[3] General Data Protection Regulation (GDPR) – Official Legal Text. (2022, 7 April). General Data Protection Regulation (GDPR). Accessed on April 13 2022, from https://gdpr-info.eu/

[4] Data protection impact assessment (DPIA). (z.d.). Autoriteit Persoonsgegevens. Acced on April 15 2022, from https://autoriteitpersoonsgegevens.nl/nl/zelf-doen/data-protection-impact-assessment-dpia#wat-zijn-de-criteria-van-de-europese-privacytoezichthouders-6668

[5] Gordon B. G. (2020). Vulnerability in Research: Basic Ethical Concepts and General Approach to Review. The Ochsner journal, 20(1), 34–38. https://doi.org/10.31486/toj.19.0079

[6] Stilgoe, J., Owen, R., & Macnaghten, P. (2020). Developing a framework for responsible innovation. In The Ethics of Nanotechnology, Geoengineering and Clean Energy (pp. 347-359). Routledge.