This detailed training guide is useful to anyone who is planning to conduct research in person, or online via video or audio conferencing. It covers topics such as how research differs from outreach, how to behave as a researcher, how to conduct a good interview, and more.
We thank the Food Systems Research Fund for supporting the creation of this guide.
Research Is Different From Other Outreach
Everyone should read this section.
Many people are familiar with best practices for interacting with the public from having done some form of outreach (e.g., in animal advocacy), public service work, or even retail. However, conducting research is different from these cases in a few important ways, and sometimes their goals may be at odds. The cardinal rule of being a researcher on the ground, interacting with participants, is to treat all participants and potential participants as similarly as possible—in everything from your manner to your appearance, to the way the study is presented.
As an example, in this study, we were interested in whether watching a video about factory farming could influence people’s animal product consumption and attitudes toward farmed animals. Some of the participants watched a video before answering our survey and others didn’t. This difference indicates the condition the participant was in. The only way we can say for sure that differences in people’s consumption behavior and attitudes were the result of watching the video is if the video is the only thing that differs from the condition with no video.
For instance, if people who watch the video eat fewer animal products than people who do not watch a video, we would like to be sure that the video is the reason for the difference. But if you, as the researcher on the ground, were wearing a t-shirt that said “Save the Animals” on the video day, and a blank t-shirt the other day, that could be the real reason for the difference, and we would have no way of knowing. That is why consistency is key.
Remember also that you are an important source of potential bias in a study. If there are differences between participants or you have conditions like in this example, you shouldn’t have one research assistant (RA) collecting data for one condition or one type of participant and another for the other condition or type. For instance, you wouldn’t want RA 1 collecting all the data from participants under 25 and RA 2 collecting all the data from participants over 25—then if you find a difference by age group you wouldn’t know if it’s due to their age or who asked them the questions.
Bottom line: If you have more than one RA, assign them to participants, days, or conditions randomly.
Read this if you will have a booth or other static space for your study.
The first step of outreach is getting the attention of the public. Make sure to pick a location with plenty of foot traffic to give yourself a higher chance of reaching a good amount of people. To ensure consistency, your location should be the same or as similar as possible every day.
If you are studying the effectiveness of an outreach method or message, you can have your usual banners or signs, as long as every participant will see the same thing–or the only thing that varies between them is what you’re studying (like the effectiveness of different messages). In general, anything that is the same across conditions, days, and people, and is part of your usual outreach is fine.
If possible, set up your booth so that when participants leave the study area they exit away from the line of those who are waiting to complete it. That way, people will not hear others comment about what they have experienced in case it may influence someone’s decision to participate or create some other undesired effect.
Consistency Across Experimental Conditions
Read this if you have multiple conditions in your study.
If you have multiple conditions in your study, it is crucial that different types of people do not end up being recruited to the different conditions. To ensure this, recruitment should be done with a script or invitation letter that is identical for everyone, regardless of condition or other factors.
In person, if your intervention is something obvious that will look different in some conditions than others (like wearing a VR headset in one condition and not another), you also need to prevent potential participants who approach the booth from seeing it, as it might result in different types of people being attracted to one condition versus another.
One option is some sort of physical screen (e.g., a tent). Another is recruiting in a separate location (e.g., across a square/around the corner), but this is less helpful as it doesn’t prevent people who come across the booth from inquiring about participation.
Read this if you are collecting data in-person or via videoconference.
The first thing someone will notice is your appearance. If you look unkempt, it is less likely that someone will consider you credible or want to approach you.
In general, we recommend that your choice of clothing reflect the kind of audience you want to attract. For example, if you are doing outreach at a college campus, dress in a way that resembles the students versus a suit and tie or other formal attire. It is important that the target audience feels like they can relate to you.
- Your clothing should be relatively plain and simple, with no slogans, brand names, or other writing visible. The exception to this is that if you are studying the effect of outreach and that outreach usually involves specific attire. In that case, wear the appropriate attire for that outreach. Appropriate clothing is important in part to avoid alienating potential participants who dislike what is written on your clothing, and in part to ensure that participants in different conditions of the study are given as consistent an experience as possible—your appearance is part of that.
- Dress comfortably and for the weather. Wearing comfortable shoes and layering clothes will help keep you on your feet ready to go.
- Body language says a lot about your attitude and approachability. Make sure to maintain good posture and a smile on your face.
Read this if you are recruiting people in person—for example, passersby or those who stop to look at your booth.
You have two goals during recruitment: to attract people to participate in your study and to ensure consistency across all conditions and participants. Both are important, but don’t sacrifice consistency to recruit more people, or your results may be meaningless.
For example, in the video study mentioned in the first section, we had a condition with a VR headset (which is exciting to most people) and a control condition with just a survey. It would have been much easier to recruit participants in the VR condition if we told them that’s what they’d be doing, but then the type of people in each condition may have differed because they’d have signed up under different information conditions. Instead, we used a generic offer that had a monetary incentive to encourage participation consistently across conditions: “Hi. Are you interested in participating in important research about food and animal welfare? You’ll be entered in a drawing to win $1,000.”
It’s okay to tweak wording to feel natural to you, but everyone should be approaching potential participants in essentially the same way. If you’re used to doing outreach, this may be difficult—it may feel like you aren’t being as effective in your outreach as a result. And that may actually be true! Just remember that conducting high-quality research is an important goal in and of itself—more important than maximizing your outreach for now because it will help you do so even better in the future.
When people ask questions before agreeing to participate…
You can provide additional information that is the same for everyone, regardless of condition. Decide what this information should be before beginning data collection. For instance:
- How long it takes to participate (e.g., “Your participation will take 10 to 15 minutes today, and we will send you a quick follow-up survey in a month, which will take an additional 2-3 minutes.”)
- What they get, if anything (e.g., a dollar, a draw entry, a coupon)
- The basics of what they’ll be asked to do, with no specifics about which condition they’re in (e.g., “Mostly it consists of some simple questions about your attitudes and the foods that you eat. Some people are asked to watch a video.”)
- How their contact information (if collected) will be used
If they ask more specific questions, you have an ethical obligation never to lie to them. Keep the information as general as possible while still in this recruitment stage so as not to bias who opts-in to the different conditions. For something like our example study, the video had graphic imagery of factory farms that may have made some participants uncomfortable. However, you should not bring this up during initial recruitment. Once they’ve expressed an interest in participating, you should tell them more about the study. At this point we gave a more in-depth explanation of the video and its graphic nature. They can still withdraw from the study if they don’t want to watch, but this prevents recruitment bias.
If a prospective participant asks other specific questions about the design of the study, you can politely decline to respond while telling them why. For example, “I cannot give you all of that information now because previous research has shown that when participants in a study know what the exact goals of the study are, it tends to influence the way they respond to the study, which makes the research useless. However, at the very end of the study, the last thing you get from us will be a short letter describing the full design of the study and what we hope to learn from it. It will also have the researchers’ contact info so that if you still have questions or concerns, you will be able to ask them directly.”
Everyone should read this section.
For any person-to-person data collection, you should have a “script”—that is, a guide for the RA to follow. Some of this is a more literal script, in the sense of what to say to participants, while other parts may include instructions that you don’t need to read aloud: when to give the participant the consent form, when to ask them to move from one task to another, what to input on a tablet before handing it to them–even who to call in case of a study emergency like running out of consent forms or incentives.
As noted in a section above, it’s generally okay to tweak the wording of a script to feel natural to you, as long as the gist of the message is the same, all key points are covered, and you treat all participants in essentially the same way.
As an example, here is the script from Faunalytics’ “Reduce” Or “Go Veg” study:
At the beginning of the day, make sure:
- You are wearing your t-shirt
- You have your script, email signup list, & pen on clipboard
- Survey loaded on tablets, headphones plugged in
- The sound is at the right volume
- You have the reward vouchers handy
“Hi, do you want to complete a video-survey and get $5 off at Rooster’s today?”
“The survey is anonymous and takes less than five minutes. It’s being run by an animal advocacy group called Faunalytics.
We’re interested in your opinions about diet and animal welfare. If you want to do it, you’ll watch a video and be asked some questions, all on the tablet. There are just a couple of screener questions first, to make sure you’re eligible to participate. If not, it will ask you to return the tablet. Otherwise, it’ll take you straight to the video.
There are a few different videos and some include footage from factory farms–nothing graphic, but you can look away if you find it distressing. This study has been reviewed by the Carleton University Research Ethics Board. If you decide you don’t want to participate, you can withdraw at any time without penalty.”
“Do you have any questions about it?”
“Do you voluntarily agree to participate?”
Give them the tablet and show them how to advance.
When survey is completed:
- Ask them to put email on clipboard list
- While they’re doing that, get a voucher
- Put date on voucher with coloured Sharpie (format: Sept 4)
- Record coupon code in Qualtrics
- Check that their email is legible.
“Thanks! You’ll receive an email with the study info at the end of the semester when we’re finished. Thank you for participating!”
- Don’t ask if they have questions
- But if they do, answer honestly but don’t give too much information
- If they ask whether we’re tracking the voucher, say “yes, but we won’t know who you are, we’ll just be able to link it to your survey answers.”
- If you don’t know the answer to something, give them my [lead researcher’s] email address: [email protected]
- If anything off-script happens, take notes about it. Participant number, date/time, what occurred. Let me know at the end of the day.
Read this if you are conducting interviews.
The script above doesn’t have many questions in it because it was for an experiment, not an interview. Interviews obviously include many more.
Research interviews have varying degrees of structure to them. Broadly, there are three levels:
- Structured interviews involve asking a series of questions exactly as they are written, without deviating from them.
- Unstructured interviews are the opposite. They are essentially conversations on a theme, with no set questions determined in advance.
- For most social science, we recommend the middle ground of a semi-structured interview. These typically involve a predetermined set of questions that is asked of all participants, but the interviewer can add questions or follow up on answers to get more detail as needed.
Being a good interviewer is similar to being a good listener. Many of the skills transfer. There are three core principles to remember, described below from the perspective of a semi-structured interview:
1. Learn From Your Interviewee
Ask your question, then listen and make sure you understand their answer. If you think of your job as “paying attention,” you can end up only engaging shallowly with the conversation, nodding along. Rather, you are here to learn about your interviewee—opinion, history, or experience—as it pertains to the topic of the interview.
If you don’t understand something in what they said, ask for clarification or more detail. For example:
- “I don’t quite know what you mean, could you give me an example?”
- “X thing that you said was interesting. Could you say a little more about that to help me understand?”
2. Be Neutral In Opinion…
Your opinion on the topics covered in the interview needs to be left at home. Ask and respond to questions in an unbiased, non-judgmental way. If you can’t do that, there’s no shame in it—some topics are emotionally loaded for many people. But if that’s the case, find another interviewer. Your results will be biased if your opinion comes through, even in nonverbal reactions to the interviewee’s responses.
That’s why the example questions in the point above are so neutral. You don’t want to accidentally lead your interviewee to give a particular answer.
That said, if you think you understand but want to make sure, at that point it is fine to paraphrase what the person said, reflecting it back to them, with the explicit request to say whether your interpretation is right or wrong. For example:
- “Just to make sure I understand, it sounds like you think pigs’ intelligence is the most important point to make in advocacy messaging. Did I get that right?”
- “So would it be fair to say that you would like to see wider use of Trap-Neuter-Return programs, or did I misunderstand?”
Let your participant think. Be silent if your interviewee takes some time to think before answering. Interrupting their thought process is another way you could bias their response. Sitting in silence can feel uncomfortable for some interviewers, so if that describes you, you may have to remind yourself to wait out the discomfort.
With all this neutrality, sometimes it can be difficult to know how to respond naturally to what a person is telling you. Point #3 below has some tips about that.
3. …But Engaged In Attitude
At the end of the day, an interview is a type of conversation and it would be nice if it felt like one. There are a few things you can do to come across as an engaged conversationalist while still maintaining your research neutrality. Some of those things are already covered above: asking follow-up questions for clarification or to ensure you understand make it clear that you care about understanding their point of view.
Additionally, do your best to remember what they said earlier in the conversation and if anything seems related or contradictory, ask about it. This isn’t hard to do if you treat the conversation as an opportunity to learn about the person—you’re just putting all the pieces together. For example:
- “I remember that earlier in the conversation you mentioned you started out in a campus Effective Altruism group. Is that the same group you’re talking about now?”
- “Oh okay, so you had fish at Thanksgiving. I remember that earlier you mentioned that you’re vegetarian, so could you say a little more about how that works for you?”
Show you’re listening with your reactions. You can use verbal and non-verbal reactions to indicate interest and understanding, or reflect the participants’ emotions. For example:
- Most people have a natural “I’m listening” response during conversations—pay attention to your next few casual conversations to find out what yours are. It’s generally fine to say whatever you normally would, whether that’s “cool” or “okay” or “gotcha” once in a while. If this isn’t something you normally do, you might want to try it out, but don’t if it feels too awkward—that will come across to participants.
- Smiling and nodding are both useful signals of attention and we often do them naturally when listening anyway.
- The exception to the neutrality rule is if you’re reflecting something the participant is expressing. If they tell you something that made them happy, you can and should be happy in your reaction. “Wow, that’s awesome!” is perfectly appropriate in this context. If they share a sad story, feel free to say “I’m sorry to hear that” before following up with another question.
It takes a little while for most people to feel comfortable as an interviewer, and it can be draining to conduct interviews, especially for more introverted people. You will make mistakes from time to time, even once you’re a confident, experienced interviewer.
Accept that mistakes will happen and forgive yourself when they do. There’s a lot to remember while also carrying on a conversation in real time, and you’re bound to forget something sooner or later.
- If you’ve skipped a question, you’re fine. Just ask it as soon as it feels natural to do so. You can ask questions out of order to better suit the flow of the conversation anyway.
- For other mistakes, make a note of it in your dataset with that participant’s other responses, or on the sheet where you took notes on their interview—wherever you will come across it when it comes time to analyze the results. At that time, the lead researcher can decide how to handle it. For example:
- Participant 3: I forgot to ask about their workplace.
- Participant 9: I accidentally made a leading statement about veganism before asking if they were vegan.
- Participant 20: I was wearing a shirt with ‘Mercy for Animals’ written on it.