Find, filter, favourite: How to select participants for AI tasks
|
26
July 2023
|
26
July 2023
|
26
July 2023
If you’ve read our blog post on finding trustworthy participants for your AI training tasks, you’ll know how important it is to recruit people who’ll give you honest, detailed input.
And you’ll be aware that a platform like Prolific does all that hard work in sourcing them for you.
But with so many eager, engaged participants available, how do you whittle them down to make sure they’re the perfect fit for your specific study?
Here are some of our top tips.
We offer over 300 free filters that allow you to be picky about who you choose.
When building your study, simply scroll to the ‘Prescreen participants’ section and add criteria participants must meet before even applying to take part in your research.
For example, you might only want to hear from people living in the UK, those within a set age range, or those with a certain level of education.
A favourite filter of our current users is ‘First Language’.
Most of our users are training English-speaking AI models, and need participants who have English as their native tongue. The ‘First Language’ filter makes findings these participants nice and easy.
A good indicator of how well a participant will perform is how they’ve done in the past.
So, you may want to target by 99% approval rate, which means the person hasn’t been rejected by other researchers.
Combine this with the number of previous submissions, which shows how many test submissions they’ve completed without being rejected, and you’re likely working with a high-quality participant. (We recommend setting this at a minimum of 200.)
Once you’ve filtered your participants down to a level you’re comfortable with, we suggest running an initial study to see who provides you with the richest data.
You can then consider these people your core group, and feel confident they’ll give you the output your AI model needs to learn and grow.
With the Prolific API, you can access extra features that further streamline your search. Prolific tends to release new features on the API first, so by using our API you’ll get early access to the latest benefits.
Find yourself using the same filters again and again? With filter sets, you can create and save specific sets of participant filters, then apply them easily each time you set up a new task.
Participant groups empower you to automatically ‘save’ a core group of people you want to work with. So you can easily find and reach out to them whenever you want. This is ideal if you’ve already trained a few participants in how to complete your tasks.
If you’re happy to let the same participants respond to your tasks repeatedly, you can also use our ‘multiple submissions allowed’ feature.
Once you’ve entered a maximum figure, each participant can complete your study that number of times before it shuts off for them.
That means you can extract as much data as possible from your highest-performing people.
To learn more about how we can help you conduct your AI research, take a look at our dedicated AI page. And when you’re ready, get in touch with our helpful team of humans.
Fresh out of YC's Summer 2019 batch, we want to share some of our most interesting learnings. If you're a startup founder or enthusiast and want to learn about product-market fit, growth experimentation and culture setting, you're in the right place!
Today Prolific is turning 5 years old – Happy Birthday to us! 🥳 It's been a remarkable journey so far. 3000+ researchers from science and industry have used Prolific last year, we have 45,000 quarterly active participants, and we've seen 200% year-on-year growth. But we're only getting started. In this post, I'll tell you a little bit about our journey, give credit where it's due!, and tell you about our exciting plans for the future.