RECRUITING PEOPLE TO TAKE PART IN YOUR RESEARCH

If you’re doing research with people and need to collect new data from them, you will need to work out how to recruit them. Sometimes, you can use existing data - and I will have a page on this coming soon.

If you do need to recruit people to take part in your study, you will need to consider a few things:
1) Who you need to recruit, e.g., specific demographics, specific behaviours, etc.
2) Whether people can take part in their own time online (e.g., an online survey or experiment), or if they can complete online or via telephone at a time that works for you (e.g., an interview), or if you need them to come to you (e.g., an in-person study)
3) How you might reward them for taking part, if at all
4) Whether you need to follow up with them.

Then we need to consider how we might recruit people. There are lots of options:
1) Recruit by getting in touch with people via where they are physically, e.g., flyers, QR codes in strategic locations
2) Recruit via snowball sampling
3) Recruit online, e.g., via social media
4) Recruit via existing research pools, e.g., market research panels like Qualtrics, Pureprofile
5) Recruit representative samples, e.g., for prevalence studies

WHO YOU NEED AND WHAT YOU NEED THEM TO DO

1) Who you need to recruit, e.g., specific demographics, specific behaviours, etc.

This is the big one. Sometimes, you might want to get a wide sample - anyone you can get your hands on. If so, it’s quite hard to get a decent sample like this. Very often, how you recruit people will lead to biases in terms of who you can get. For example, older people spend less time online, so if you have an online method of recruiting people, then you’ll find that you get younger people and not older people - your sample will be biased. How will you deal with that?

Very often, we’re after specific people - for example, people who take part in a certain behaviour (e.g., who bet on sports), or fit certain demographic criteria (e.g., men aged 18-29). If your criteria are too narrow (e.g., you want men who live in a specific town who are aged exactly 18 and who bet on sports every single day), you’re going to find recruitment very difficult because there just aren’t that many people who fit into that category.

You might also have exclusion criteria. Inclusion criteria are one thing (e.g., men aged 18-29) who bet online at least once a month, but you might want to exclude them if they only bet on a specific product that you’re not interested in, say esports.

Look at striking a balance between going too broad and too narrow. Make sure that your criteria are appropriate for your study, of course! But, when you’re planning your study, you need to think carefully about whether your criteria are achievable.

2) Whether people can take part in their own time online (e.g., an online survey or experiment), or if they can complete online or via telephone at a time that works for you (e.g., an interview), or if you need them to come to you (e.g., an in-person study)

What are you asking people to do?

If you want them to come to an in-person study, then you’ll need to work out which times will suit your target audience best, and probably run a few sessions. You may need to help them cover their travel, possibly work out insurance, and ensure you have facilities (e.g., bathrooms) and maybe even refreshments. For example, one of my students soon-to-be-Dr Annie Nolan ran a study where participants attended a tree-planting session, and they had to attend in person to go to a particular location to plant the trees. She arranged transport for them (e.g., buses from a particular location), refreshments, etc. There’s a lot to consider! Or, you might ask them to come to your lab so you can take samples, or do whatever you need to do to them (ethically-speaking of course). Naturally, it’s harder to recruit people for this sort of stuff.

If you need them to take part in an interview, sometimes this can be done in-person, but more often these days it is done via a telephone call or online interview via Zoom or similar. However, you still need them to attend at a time when you (or another interviewer) can attend, so this often requires you to be available at a range of times. That’s fine, interviews are great because you get the opportunity to ask them questions and ask for more detail about their answers. You just have to work in with when they are available. There are also some ethical issues around recording people, so work that through with your ethics committee.

The easiest for you, and often for people, is that they complete something in their own time, say an online survey or an online experiment that doesn’t require you to be present when they do it. This is very often the easiest to get people to do, and is also a lot easier for you. But, because you don’t meet them, sometimes it’s a bit unclear if they are legitimate responses. See my page on protecting your survey from bots and duplicates, especially if you’re rewarding people for taking part e.g., by paying them or by running a prize draw.

3) How you might reward them for taking part, if at all

You don’t have to reward people for taking part in your research, but often you’ll find you won’t get many people if you don’t It will depend a bit on whether your project is a bit more naturally attractive or not. For example, with my PhD, I had people tasting wine samples, and it wasn’t hard to recruit people! But something like an online survey, if done in a boring way, won’t get many people through. You can make them more attractive, e.g., through interactive surveys that give people feedback.

People will typically reward people either by paying each person who takes part (this is what market research panels tend to do), or by putting people into a prize draw and selecting one or more winners. Both require a budget. The benefit of paying everyone is that it’s more attractive, but you’ll most likely want to cap numbers to make sure that you don’t blow your budget. If you had budgeted for 200 people at $5 per head, and have promised everyone $5 per head, but you get 1,000 people, what are you going to do? Note you can cap numbers via quotas in Qualtrics.

If you run a prize draw, you’re at far less risk of going over budget, but it may be less attractive. It’s common to use these. You may find you want to talk to the ethics committee about running this.

If you do want to reward people, and especially if you’re advertising online, you’ll find that you’ll get hit by bots - automated responses that aim to take the survey lots of times to rip you off. I have more information on this page, as well as some measures to help minimise the impact of bots.

4) Whether you need to follow up with them.

If you want to follow up with people, say a multi-wave study, you’ll generally need to capture their contact details, or have a panel manage re-contact. You will lose lots of people over time through natural attrition, so build this in to your thinking about how many people you need to recruit initially. Attrition will depend on factors like how much you’re offering them to take part in each wave of data collection, and how long the gap is between waves.

If you want to get their contact information to manage it yourself, that’s generally fine if you recruit them yourself, e.g., via social media advertising. But, if you’re recruiting via a market research panel, most won’t want to share their information with you, or allow you to collect it from them. This is because this is how they make their money.

WAYS TO RECRUIT PEOPLE

1) Recruit by getting in touch with people via where they are physically, e.g., flyers, QR codes in strategic locations

You can recruit people via physical locations, such as by handing out flyers in shops, or posting notices on relevant billboards. As always, try to make sure that these are attractive for the relevant audience, and also that they are located in the places where the people you are after will hang out. If you’re running a study on helping people improve their fitness, you might consider something like a gym. If you want people who take part in team sport, you might consider flyers or QR codes at your local sporting club locations. Of course, make sure you work with people at these places to get permission, and to help you spread the word.

Sometimes there are community noticeboards around, e.g., in cafes or libraries. These can be useful too.

Keep a track of where you have put them, and when the study is over, it’s common courtesy to go around and clean them up.

2) Recruit via snowball sampling

You can always ask people who have taken part in the study to refer their friends. For example, you could have a link at the end of a survey that people can share. You might even incentivise them - work out a way to determine who has referred the most people, and have a prize for that.

Be aware with snowball samples that sometimes this leads to samples of people who all know each other, which can cause a bit of bias. Depending on the study, there may also be privacy issues. For example, if the study is specifically on people with a certain condition, then asking someone to spread the study may give away the fact that they have that condition.

3) Recruit online, e.g., via social media

I haven’t had a chance to write this bit yet - will hopefully do it soon. Main things to consider:
- Usually, you’re after particular people for research, such as people who take part in a particular behaviour, or who fit a certain profile (e.g., parents). Look for social media groups that cater to these interests.
- You will need to show your social media ads to an ethics committee. Have them developed before you post them. Keep them short and snappy, and where possible make it clear who the research is for. For example, “This study is open to women aged 18-30 from Australia”. You may sometimes need to balance this with giving too much away about who is eligible - sometimes if you tell people too much, you’ll get people lying about who they are, or their behaviour, in order to make it into the survey. Work through this issue in your ethics application.
- You may want to set up a new profile, rather than use your personal profile. Just so you don’t get weird people adding you as friends, or contacting you.
- Always check with admins or owners of social media groups before posting. Be sure to check any rules for groups, too.
- You can pay to boost posts on certain social media platforms, e.g., Meta (Facebook/Instagram), Twitter/X, etc. If it’s within your budget, by all means consider this.
- If you’re using a prize draw, or paying participants, you will get a lot of “bots”. Be sure to build bot protection into your study, such as through online surveys. I have a page on bot protection here.
- You usually can’t just post once and then expect responses to keep rolling in. You will need to repost or similar. There is a balance here with annoying everyone in the group, so make sure that you don’t push it too far!
- Occasionally, there have been issues, e.g., running into people who are complaining about what you’re doing. Be sure to work with your corporate communications team if you’re from a uni if this happens. CQU’s Corporate Comms team is great.
- Give back to the community when the research is done, if appropriate. Provide a summary of what you found (no identifying results in there please).
- Make sure you take down any posts once the study is over.

4) Recruit via existing research pools, e.g., market research panels like Qualtrics, Pureprofile

If you have a research budget, there are existing market research pools that you can pay to use. Honours students usually won’t have a budget, RHD (Masters, PhD) may have a bit of a budget. If you’re applying for a grant, you can build recruitment via panels into your budget.

I tend to use two main market research panels: Qualtrics and Pureprofile.

Qualtrics
When most people think Qualtrics, we think of the online survey platform. If you’re using Qualtrics in your research, you might be interested in my Qualtrics resources that show some tips and tricks in using the platform.

Qualtrics also offers help with recruiting people. This is a paid service, so you will need a budget. Qualtrics does not own their own research panels, but they recruit via other existing panels. They manage the recruitment process so you don’t need to worry, including helping you set up your survey, checking data quality etc. Additional paid services include things like translating surveys into other languages.

To line up recruitment with Qualtrics, you will need to contact your account manager. At CQU, this is Christopher Arditti christophera@qualtrics.com . It may be another person if you’re at another university, but Christopher should be able to put you in contact with them. Tell them I said hi.

Christopher helpfully sent me two documents. These are current as of 2025. One is a series of PowerPoint slides that talks through their services at a high level. The other answers 28 questions about buying online samples from Qualtrics.

Qualtrics usually has a minimum spend, somewhere around AU$4,000. This may rule out some people with very minimal budgets, or who are looking for a smaller recruitment.

Pureprofile
Pureprofile owns their own panel, but also partners with other panels as needed. They don’t have a minimum amount, usually, and I’ve found they’re very happy to work on most projects. Sometimes, Qualtrics has declined to recruit for a project.

My contact at Pureprofile is Ivan Chen ivan.chen@pureprofile.com . I’ve worked with Ivan for years and he’s great. Tell him I sent you. Pureprofile doesn’t do everything by default - they will send you the links to include in your survey to make their recruitment work, and won’t check data quality automatically. However, you can ask for these services to be added, and they are are generally pretty cheap - just a few hundred dollars. Worth asking Ivan about these things.

Ivan has also sent a list of 37 questions to help people buy online samples from Pureprofile.

To get a quote from Qualtrics or Pureprofile, you will need to tell them:
- How many people you’re hoping to get. You can ask for a range, e.g., quote for 500 and 1,000. You might also ask them their maximum feasible number - sometimes, they’ll only be able to provide a certain number, especially if your criteria are too narrow
- When the project will be in-field
- Any inclusion or exclusion criteria (e.g., 18+, from New South Wales only, who bet on sports at least monthly). Note that sometimes they will have this info in the user profiles (e.g., age, location), but sometimes you’ll need to ask questions in the survey (e.g., gambling behaviour)
- Any quotas (e.g., no more than 70% men)
- Approximate survey length (in minutes), both on average and for the longest possible survey (e.g., if there are lots of questions only answered by some people)
- If there are specific criteria, e.g., the example of betting on sports at least monthly, they will want to know an approximate prevalence, i.e., what percentage of people in NSW aged 18+ bet on sports. This is often an educated guess, but it often is the most important factor behind the price.
- Any other information. Sometimes this might include things like wanting to follow people up, or wanting to capture their contact information (aka personally identifying information or PII). PII is often frowned upon, because these panels make their money by giving you contact with these people. If they give away all their contact info, they don’t have a product anymore.

Note that there are also plenty of other research panels around. I have explored plenty of them over time, and have had some good experiences with others (and some not-so-good experiences with some too). Don’t feel limited in terms of using Qualtrics or Pureprofile for recruitment - others may be better for particular things, e.g., qual recruitment.

5) Recruit representative samples, e.g., for prevalence studies

Representative samples are really important for things like prevalence studies. You don’t always need them for research, but when you do need them, they’re really important. See this paper for a discussion on this.

Recruiting representative samples is difficult. A representative sample doesn’t just mean one where there are similar demographic splits to the population. For example, you might be able to get a sample that is approximately the same age and gender splits as what we see in the Australian population, but that doesn’t mean that it’s representative. The idea is that as many people in the population have the opportunity to be selected. One of the most common ways to do this is via calling random telephone numbers for a telephone survey. Of course, lots of people hang up and don’t answer, but they at least had an opportunity to be included. In contrast, people who sign up for market research panels are people who have agreed to do that - and often this is a specific type of person, say who needs a bit more money and wants to be paid via surveys, or who has the time to do it. They’re usually nto considered to be representative.

We’ve worked with a few companies who are fantastic with conducting telephone samples like this. Recently, we’ve done great work with Ipsos and I would recommend them for this work. Another is ANU’s Social Research Centre.

These samples typically cost a LOT of money, hundreds of thousands of dollars, because there’s a lot of work involved - lots of people on telephone lines making lots of calls. The results are fantastic, and they are the current best practice on prevalence surveys.

Other ways that this can be done is through things like door knocking (especially if the population is small, e.g., for a particular town only), or mailouts. There are some ways to push people to complete online once you’ve initially recruited them, too. This space continues to evolve, especially as telephone response rates continue to drop.