Citizen science is a fascinating thing, both because of the interesting projects undertaken, but also the way participants work with their peers.  In many projects, participants remain strangers to one another, whereas in others the bond is much closer.

A recent paper set out to examine just what impact peer pressure had on the contributions of citizen scientists.  The study saw a model created of human behavior that tried to predict how we behave in certain circumstances.  The authors believe that the model could be crucial in helping to encourage participation in science.

At the heart of the study was the Brooklyn Atlantis project, whereby water drones are used to patrol the Gowanus Canal in New York, and report back data on the quality of the water, and images from above and below the waterline.  This data is analyzed by a team of citizen scientists, with participants tagging images with any objects contained within them.

Unequal distribution

It’s well known that the bulk of work undertaken by citizen scientists is done by a relatively small number of diehard supporters, and so increasing, and widening participation is a goal for most projects.

The researchers thought that peer pressure may be a useful prod, even if it was only a virtual peer.  The interface to the Brooklyn Atlantis project was modified so that the performance of a virtual peer was given a prime position.  There were five different levels of performance shown to participants to try and discover what impact each would have.

Participants were split into four groups.  A control group engaged with the standard interface (ie with no social comparison) and the remaining four received peer information whereby the peer outperformed, underperformed or was on a par with the user.

Peer support

Perhaps unsurprisingly, the results showed that our peers do influence our performance, especially when we think they are outperforming us.  The best performers in the experiment were consistently those who were led to believe that their peers were doing better than them.

The inverse was also the case, with the worst performing participants consistently being those who were led to believe that they were doing better than their peers.

“Social comparison is a strong driver of behavior, and it’s exciting to see that even simulated performance was enough to influence our participants to tag more or fewer objects. Even more exciting was the fact that we can anticipate such a response using a mathematical model,” the authors say.

Establishing the norm

The authors contend that we generally try and reflect what we believe to be the standard behavior in any group, and therefore setting a slightly higher norm can be an easy and effective way of boosting performance.

“The study taught us how the design of a social participation system can benefit from incorporating social psychology research,” they explain.

Citizen science is a growing field, but I think these findings will have relevance for any project that is trying to elicit wide participation as it highlights the power peer performance can have on our motivation levels.