Seven tips for managing large sample sizes in short term online communities


Online communities have many benefits. While not appropriate for every audience or topic, they can make it easier to reach geographically dispersed groups, gather real-time, multi-media data, and be more cost effective than running comparable research face-to-face. They do, however, have their challenges, particularly when large numbers of participants are involved. In this post Anna Cordes (Senior Policy Researcher, Which?) reflects on her experiences delivering online communities and what researchers should keep in mind when designing online studies with large numbers of participants.   


I love online communities. Although depth interviews have their charm, and focus group wrangling can bring a buzz, the opportunity to bring people together across continents is a thing unto itself. That you can iteratively build insights, revisit themes as they emerge and give both researchers and participants an opportunity to reflect, is an added bonus.

In their broadest sense, online communities are online spaces where people come together to discuss a particular topic. Some develop naturally (think about the types of communities which coalesce on social media platforms, around particular hashtags or in below-the-line comments on articles); some are curated by brands to engage their audiences - for instance, discussion boards or conversation pages; and others are commissioned for research purposes. It’s this last type of community we are focusing on here. 

Online research communities generally have specific objectives and research questions, are ‘open’ for a fixed period of time and have a pre-determined number of participants. That said they can come in a range of shapes and sizes:

From small to large

Although online communities commonly have 20-30 participants, online community platforms generally give you a large degree of flexibility over how many people you wish to take part and for how long. While working in research agencies and more recently client-side, I’ve been involved with micro studies lasting just a few days and involving around a dozen participants, through to large-scale studies crossing several markets and involving over a hundred participants. As a rule of thumb, larger online communities are better when you want to gather insights at scale and cover or compare experiences in different locations. 

Variety of tasks and activities 

You can also ask participants to complete a variety of different types of activities. Research platforms have moved beyond offering just static message boards. They can give you a high degree of control over the types of question you want to ask and the format of response you’d like to receive back.  Most let you: 
 - Have public or private tasks 
 - Decide whether or not (and when) you want to socialise participants’ responses 
 - Ask closed and open questions 
 - Choose what format you would like participants to reply in (text, video, audio, or pictures) and how long you would like replies to be (for instance, setting a minimum or maximum number of photo submissions, or setting a word count)
 - Segment participants and code responses in real time 

The challenges of online communities 

While online research communities can be a valuable tool for researchers, they do come with their challenges. They should not be seen as a ‘set it and forget it’ option for research. Ensuring participant engagement, and that research objectives are being met, relies on skilled moderation, planning and forethought.

While challenges exist when running online research communities of any size, they can compound as you increase participant numbers. Once you breach the 30 participant mark, the volume of data can quickly become unmanageable for a single researcher. It is similarly challenging to ensure each participant is being moderated and their responses probed. However, running larger scale qualitative online research communities does have a range of benefits, including the insights you are able to derive. 

When designing or commissioning an online community with larger numbers of participants, consider how you will manage attrition, support analysis and help participant engagement. Below are some questions to ask yourseful and things to think about that may help you run your community:

1.  Will your participants be able to engage online? Online communities require access to the internet, either via a PC, laptop or smartphone. This can present a barrier to some participants. 

2.  Even if participants can get online, language and literacy barriers can mean it’s difficult for people to contribute fully. Some may struggle to engage with written instructions or stimuli in a less moderated online community or to communicate their views in writing. This can be overcome by having multiple ways for participants to contribute – voice memos or videos might be easier for people than typing out wordy answers. 

3.  What is your moderation plan? Managing lots of participants can be challenging. Before you get started it’s worth thinking about whether there are any sub-groups within your sample which you want to analyse.. If you’ve got more than one researcher on your project, you could assign them to look after a particular cohort and probe the subgroup’s responses. Chunking up larger samples into more manageable groups can help make sure no one group of participants is ignored and means researchers can be held accountable for how much attention they are giving to each participant. This can also help reduce attrition.

4.  How are you going to keep people engaged? Sometimes, participants drop out of the research process – it’s the way of the world. Effective moderation – asking appropriate follow up questions, for instance – is probably the best way of keeping people engaged and stops them feeling they’re just typing into the void. 

5.  Having a mix of different task types also helps. It stops people getting bored, is more inclusive and can yield richer data. For larger samples consider whether the platform you are using  can automatically provide audio or video transcriptions – this functionality can save time and support analysis.

6.  Do you need depth or breadth of insight? It is tempting to have all your participants completing all of your tasks. However, having large numbers of participants gives you the flexibility to split your sample and test more things. For instance, you may want to split your sample in half and run certain parts of the online community in parallel. Alternatively, you might want to have special tasks that only ‘unlock’ for participants which fulfil certain criteria (e.g. have had a hospital appointment in the last month, or have switched their utility provider). This is not only respectful of participants’ time, but also helps the analysis process.

7.  How will you manage your analysis and reporting? Lots of participants means lots of data! Although for some researchers and on some project timelines this will not, in itself, be a problem, those on tight-deadlines or in a commercial setting should think about how they could effectively triage responses in advance. Using a combination of closed and open questions can be an effective way of grouping participants. Many platforms let you assign participants to segments or ‘tag’ responses as you moderate and then export data based on these codes.  Similarly, some tasks, such as creating a heat map based on participants’ marking up of an image, can act as powerful outputs in themselves.

This is far from a definitive list of ways to manage large online communities. What would your top suggestions be for managing large samples? Just comment below!

AUTHOR: Anna Cordes is a Senior Policy Researcher at Which? where she works across quantitative and qualitative research projects. Before joining Which?, Anna worked as a Research Manager for Kantar Public in their qualitative team delivering projects for clients such as HMRC, the Health and Safety Executive and the Department for Business, Energy and Industrial Strategy. CONTACT: / @AVCordes