The Blog.

Make it pretty! Use styling to make your experiment relevant and prevent it from standing out in a negative way.

For my previous study, with minimal recruitment, I had a total of 290 participants from a very specific population, in just a few days of open recruitment. The completion rate was at over 56%, which is amazing when you consider that the 20-30 minutes of participation consisted of tasks that require constant vigilance and mental effort and were really no walk in the park, and definitely not entertaining. Despite of being meant to assess cognitive skills and team cohesion in League of Legends players, the experiment served another purpose - to assess the rate of engagement in what I like to refer to as native experiments.

In June, I was in contact with an ad tech company, called Celtra, about product analytics. We met a few times and once they told me that one of the very successful features they have had are their so-called native ads. The word "native" refers to the content's coherence with other media on the platform, and in the case of Celtra, who have a product for designing mobile display advertisements, this meant some relatively typical mobile advertisements introduced and displayed in a way that is fluid with the context and unintrusive to the user. For instance, in an Angry Birds game, they would have a flock of birds fly in and spread out to introduce a banner ad. In the Angry Birds Star Wars edition, they had a similar effect introduced by a light saber. It had been assessed that these ads were much less intrusive than normal mobile ads without the native component, and an increased level of user engagement was observed.

I thought this was great! I could use some of the features of the native principle in my experimental battery. And what better way to do it than to spice it up with some game art and humor. If I have an entire experimental battery meant for League of Legends players, then the battery - or at least the instruction screens and surveys where cognitive skills were not of essence - should match that goal! So I colored up the test battery and got tons of positive feedback - some even told me that it was the only test battery where they had ever actually enjoyed the participation. An encouraging word here, some bright colors there, a piece of game art or two - and I was getting positive feedback for the design, which - despite of all the effort I put in - would have made any graphic designer cringe, or perhaps vomit. Here I have to confess that even if I'm very happy with my skills when it comes to experimental design, analytics and basic web development, I am no designer.

Put effort into the design. I understand that most experiments rely on a black-and-white style with your default RGB colors. Clearly cut shapes, bitmap images and strenuous repetition until you reach the "thank you for your time"-message on the last page of the experiment. This works fine if all of your participants are undergraduate students motivated by the thought of course credit, but if you're recruiting participants online, you will have to do much, much better.

Why do you think Lumosity is so popular? It attracts your curiosity, your engagement and even talks to your sense of aesthetics. Their Flanker task consists of an arrangement of birds against the blue sky. Yours probably has screamingly white triangles against a pitch black background. But good design doesn't only exist to elicit curiosity. It is also great for engaging difficult audiences, such as children. All you need is a fun set of instructions, some colors and a pinch of gamification, and you might not only see a difference in engagement, but also in performance and adherence.

Finally, it should be said that there is only so much you can do with style or with external incentive - the two things I've mentioned as motivators throughout this series of blog posts. Style works in interaction with contact, feedback, encouragement, convenience of participation - and all the other factors of the experiment. And in a way, style is still a hygienic factor, as it is something you use to guarantee that the experiment is not intrusive and does not negatively stand out from its context.